feat(bin): merge GitLab merge requests through the guarded PR merge path - #2779
Merged
kunchenguid merged 5 commits intoAug 22, 2026
Merged
Conversation
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve.
… cases host-independent
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
Confidence Score: 5/5The PR appears safe to merge. No blocking failure remains. Reviews (2): Last reviewed commit: "no-mistakes: apply CI fixes" | Re-trigger Greptile |
Owner
|
Speaking as Kun's firstmate: Default-behavior for the guarded merge helper: a well-formed GitLab merge-request URL no longer refuses; it merges through VISION:
Security: no. Overlap: CI including no-mistakes is green (Require no-mistakes, Lint, Test coverage guard, portable shards, Herdr, macOS Bash snapshot, repo invariants). mergeStateStatus CLEAN, 5 ahead / 2 behind. Greptile 5/5. Bundled Merge-eligible: YES, with the #2660 overlap as a landing-order constraint. Not waiting on the captain. |
Owner
|
Speaking as Kun's firstmate: GitLab merge requests now take the same guarded merge path GitHub already has. GitHub invocations stay squash-default and byte-identical. GitLab is URL-dispatched: live VISION: fleet outlives any vendor; scripts own the refusals; GitHub path is not a new default. Class: opt-in capability (GitLab URL). CI green including no-mistakes. Landing this. |
kunchenguid
approved these changes
Aug 22, 2026
kunchenguid
left a comment
kunchenguid
left a comment
Owner
There was a problem hiding this comment.
Speaking as Kun's firstmate: guarded GitLab merge path, GitHub unchanged, green CI. Approving to land.
Owner
|
Speaking as Kun's firstmate: this is merged. Thank you @Inthuson — really appreciate you taking the time on this. |
This was referenced Aug 22, 2026
prandelicious
added a commit
to prandelicious/firstmate
that referenced
this pull request
Aug 22, 2026
* ci: gate GitHub workflows with pinned actionlint (#2517)
* fix(lint): catch malformed GitHub workflows before merge
A self-broken ci.yml cannot report its own breakage, so parse every
workflow in the local lint path that no-mistakes already runs.
* fix(lint): pin actionlint instead of Ruby for workflow lint
A self-broken ci.yml still has to fail in the local lint path, and the
named tool for that gate is actionlint, not a new Ruby runtime.
* no-mistakes(document): Clarify pinned workflow lint documentation
* fix: install pinned lint tools across supported platforms (#2546)
* fix: install pinned shellcheck and actionlint on macOS and linux arm64
The installers were hardcoded to linux amd64 and sha256sum, so a Mac
dev could not satisfy the refuse-on-mismatch lint gate. Select the
official per-platform archive and checksum, and fall back to shasum -a 256.
* no-mistakes(document): Document cross-platform pinned lint installers
* docs: reconcile test-evidence docs with store_in_repo: true (#2548)
.no-mistakes.yaml has set test.evidence.store_in_repo: true since #2355, but
CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described
the old policy of keeping evidence out of the repo in a temp directory.
The current no-mistakes behavior for store_in_repo: true is to publish each run's
test evidence to the orphan no-mistakes/evidence branch and link it from the PR
body. That branch shares no history with code branches, so evidence never enters
a pushed feature branch or the default branch, and CI's tracked personal fleet
paths rule stays accurate.
Docs only. No change to .no-mistakes.yaml or any workflow.
* docs: clarify test evidence branch storage (#2549)
* docs: correct test evidence storage comment in .no-mistakes.yaml
* no-mistakes: apply CI fixes
* docs: hint that live scouts may host their own Lavish review loop (#2563)
Make that a first-class option in always-loaded instructions so firstmate does not default to mediating and tearing the scout down between iteration rounds.
* fix(bin): report remote secondmate delivery and state truthfully (#2570)
* fix(bin): report remote secondmate delivery and state truthfully
A steer to a remote secondmate crosses fm-on.sh to a host-local fm-send
leg whose unconfirmed submit read-back (verdict=pending, typically a busy
mate whose harness queues the steer) was flattened into exit 1, so the
parent printed "error: text not submitted" / "error: text not sent" and
discarded the pending-reply expectation for a steer that had actually
landed. fm-send now carries the verdict across the ssh boundary as a
documented delivered-unconfirmed exit 3: the parent reports the steer as
delivered with confirmation pending, exits 0, keeps the expectation armed
(awaiting_report), and closes --resolve-key decisions, while transport
loss (ssh 255) and real remote failures keep failing loudly with the
remote leg's stderr attached. A local unconfirmed submit now also exits 3
with an honest non-error message and still never closes a decision key.
fm-crew-state.sh and fm-peek.sh no longer read a remote mate's endpoint
through local probes (which misreported a healthy mate as "worktree gone"
/ "can't find session: remote"): both now use the true remote source over
fm-on.sh, and an unreachable or unreadable remote reads as unknown-remote,
never as gone or dead.
* no-mistakes(document): Document remote delivery and state truth
* no-mistakes: apply CI fixes
* feat: adopt spendPriority for quota dispatch (#2574)
* Adopt quota-axi 0.1.29 spendPriority-primary array dispatch.
quota-axi 0.1.29 publishes schema 5 with selection.spendPriority as the primary comparative signal and demotes derivation fields out of default --json. Rank comparable-fit candidates on that scalar, keep runway versus the completion horizon as a hard gate, and raise the compatibility floor so a pre-consolidation build cannot reach dispatch intake.
* no-mistakes(review): Correct schema fixtures and remove prescriptive selection prompts
* no-mistakes(document): Correct quota verification evidence chronology
* Collapse quota-array-dispatch onto TOON-first spendPriority ranking.
Decide from quota-axi's default TOON; keep --json as a rare defensive fallback.
Rank by spendPriority after eligibility, reasoning-class, and runway-feasibility gates, and drop the hand-computed Pareto, pace, reserve, and window-id layers.
* no-mistakes(review): Permit ambiguous JSON fallback and correct reset fixtures
* no-mistakes(review): Correct runway semantics and escalate unresolved uncertainty
* no-mistakes(document): Document TOON-first quota dispatch evidence
* docs: add GROK_BOT.md Grok Bot system prompt (#2590)
* docs: add GROK_BOT.md Grok Bot system prompt
* docs: amend GROK_BOT.md with charter report-back and delegation marker
* docs: classify GROK_BOT.md as public-product
* docs: make GROK_BOT.md the plain Grok Bot system prompt
* docs: update GROK_BOT.md nautical terms and self-improvement (#2592)
* doc: Update language in GROK_BOT.md for clarity
Refine language for clarity and consistency in instructions.
* fix(bin): preserve inactive reconciliation scan progress (#2595)
* fix(bin): guarantee inactive-reconcile scan progress under second quantization
The inactive-outcome scan computed its aggregate deadline in whole seconds,
so a 1-second budget's effective value lands anywhere in (0,1]; a scan
starting just before a wall-clock second boundary rounded its whole budget
away mid-scan and exited having visited no child, while the durable cursor
had already advanced past the never-examined child. This is the CI flake
behind tests/fm-inactive-reconcile.test.sh's 'next bounded scan did not
resume with the following child' (watcher-wake-lock family, portable
serial 2, seen on the PR #2590 run).
Every scan now visits at least its first due child with the per-child
state-read bound floored at one second, so no invocation can be a zero-work
no-op. The outer process-group kill moves to budget+1s: the scan's own
deadline enforces the budget, and the kill is a backstop for a scan wedged
in an unbounded wait instead of a racer that routinely preempts the clean
bounded exit. The wake-lock-wait test bound tracks the backstop (3s -> 4s);
the previously flaky assertion is unchanged.
* no-mistakes(document): Document inactive-reconcile deadline backstop
* doc: Revise Firstmate delegation and communication guidelines
Refactor the guidelines for Firstmate's role and delegation process, emphasizing the importance of crewmates and asynchronous work.
* doc: Update work delegation and secret management instructions
Clarified guidelines for handing off work to crewmates and managing secrets.
* test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617)
Three assertions in tests/fm-procevent.test.sh depended on a detached runner
having finished work that the command starting it does not wait for.
reconcile's replacement runner is started through detach_runner, which only
forks: reconcile returns and counts the start before that runner has claimed
its source or exec'd its child. Any assertion taken straight after reconcile
therefore samples a race.
- The publish-before-apply recovery section left its always-ready /bin/echo
source registered across the recovery reconcile, so that reconcile launched
a competing detached poll (observed: started=1) that then raced every later
assertion for the source claim, the next capture sequence, and this home's
applied record, and outlived the section holding a live claim. It is now
retired before that reconcile - re-announcement is proven from the durable
inbox alone and needs no registration - and started=0 is asserted so a
competing poll cannot be reintroduced unnoticed. This is the same
retire-before-reconcile discipline the self-announcing section already
carries; that section acquired it after the identical race made its
"not-autohandled: self-src" assertion read "already owned: self-src".
- The crashed-leader replacement section snapshotted the replacement's claim
file and execution log behind a fixed 0.5s settle window. On a loaded
machine that window expires first, which is the CI flake behind "a
replacement runner started without recording its own claim" and "reconcile
did not start exactly one replacement source". Both effects are now waited
for with the suite's bounded wait helpers; the exact one-replacement count
is still asserted afterwards, unchanged.
- The duplicate-start section slept 0.5s for reconcile's runner to record
ownership before asserting that a second start loses to it. It now waits
for that claim.
Also tighten one assertion that could not fail as written: "autohandled:
self-src" is a substring of "not-autohandled: self-src", so the applied path
was accepted even when the runner reported the capture left for the handler.
Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced
6 failing runs, all in the crashed-leader section. On the fixed suite, 216
full runs under the same load produced none. Reverting the self-announcing
section's retire-before-reconcile line reproduces "already owned: self-src"
on the first iteration, confirming the shared mechanism.
* fix: preserve pending replies and defer remote reposts (#2618)
* fix(bin): keep pending-reply expectations honest on both send legs
Two related asymmetries let the parent-owned secondmate reply guard drop or
nag requests it should not have.
Local delivered-unconfirmed dropped the expectation. A marked request whose
submit read-back stayed unconfirmed (verdict=pending) is the same
not-a-failure outcome the remote leg reports as delivered, but fm-send
discarded the parent's pending-reply record for it, so a request that very
likely landed stopped being tracked entirely. The record now stays armed on
its unconfirmed-delivery marker: a correlated report still resolves it, and
an unanswered one still surfaces through the library's own reconciliation.
Exit 3 and the local rule that an unconfirmed answer never closes a decision
key are unchanged.
Remote replies were nagged for a repost they did not need. A remote mate's
report reaches the parent's status log only through the asynchronous mirror
in fm-procevent-remote-reply.sh, yet the guard read an absent correlated
line as proof the mate never reported - even while the answer was still in
flight, which is the common case because the mirror's poll window is
comparable to the recovery grace. The mirror now publishes one caught-up
watermark from a quiet window, and the guard admits a missing report as
evidence only once that watermark passes the turn that should have produced
it. A genuinely missed report still gets exactly one repost, and a channel
that is behind, unarmed, or broken leaves the request durably open and
un-nagged rather than nagging blind; the mirror escalates its own continuity
failures as before.
Tests: a local unconfirmed secondmate send keeps its expectation armed and
resolvable; a mirrored correlated remote reply resolves with no repost; a
stale or absent watermark withholds the repost while a fresh one still
releases it; a quiet remote window publishes the watermark and retirement
clears it.
* no-mistakes(review): Distinguish preempted polls from quiet windows
* no-mistakes(document): Clarify remote reply channel freshness
* no-mistakes(lint): Annotate shared remote preemption exit constant
* fix(bin): honor declared pauses in busy-pane wedge checks (#2619)
* fix(watch): honor a declared pause on a busy pane's completed-turn bound
A worker that declares an external wait (`paused:`) and then blocks in one
long foreground call - a review-hosting scout parked in a single blocking
`lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane
BUSY, so the stale path that already honors declared pauses never ran for it.
The busy-pane completed-turn bound instead routed it straight into
wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past
the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long
as the review stayed open.
busy_turn_bound_check now owns which absorber takes a crossed bound: a crew
whose own last status line declares an external wait or a verified captain-held
transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else
keeps the unchanged wedge timer. The discriminator is the declaration together
with liveness (the caller has already confirmed the pane is busy), never a
blanket silencing - a crew that declared nothing, or whose pane is not live,
escalates exactly as before, and a declared pause still re-surfaces once per
long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched:
the daemon owns pause triage there and already reads the same vocabulary.
The two call sites also no longer clear pause bookkeeping in the same poll the
pause cadence recorded it, which would have erased the re-surface throttle and
turned the long cadence back into a per-poll re-surface.
Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence
recheck, and the restored wedge escalation once the declaration is lifted on the
same busy over-age pane.
Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded
machine: fixed liveness budgets were reaping watchers mid-startup, so assertions
on post-poll state passed vacuously or failed spuriously. Waits that describe a
poll's outcome now wait for a completed poll cycle via the liveness beacon, the
heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit
budget is the uniform 10s already used elsewhere in the file.
* no-mistakes(review): Fail poll-cycle waits on timeout
* no-mistakes(review): Prevent poll timeout test hangs
* no-mistakes(document): Clarify paused busy-pane supervision
* doc: Enhance communication guidelines for decision-making
Added guidelines for decision communication to the captain.
* doc: Update task delegation and communication guidelines
Clarify communication protocols with crewmates regarding task delegation and reporting.
* fix(bin): reliably confirm herdr steer submission (#2647)
* fix(herdr): confirm local steers that native agent-state misses
Herdr can leave agent_status idle for a landed Claude turn and can keep
queued Enter text visible while busy, so fm-send was reporting false
swallows. Confirm those cases through the shared queued-Enter verdict
and a cleared composer, and keep a genuine idle pending composer as
unconfirmed.
* no-mistakes(review): Stop Herdr Enter retries on unreadable composers
* no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail
* no-mistakes(review): Prevent confirmation after failed Herdr Enter
* no-mistakes(review): Pace Herdr retries and clarify submit fallback
* no-mistakes(review): Align Herdr submit docs with idle fallback
* no-mistakes(document): Correct Herdr submit-confirmation documentation
* feat(bearings): add interactive Lavish fleet board (#2659)
* feat(bin): accept any-origin decision bindings with full-identity keys
An aggregation surface (the bearings board) carries captain answers for holds
across origins, but a binding was one-origin-per-source and the Lavish adapter
capped question keys at 64 chars while real full hold identities measure 69-81.
- fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker;
binding prints it verbatim and answers accepts it, so the runner's feed seam
carries an any-origin source with no runner change. In any-origin mode each
key is a full hold identity <origin>-decision-<key>, split at its first
-decision-; a key with no separator (merge/dispatch instructions) is skipped
and feeds nothing, keeping non-decision answers out of the hold ledger by
construction. Every existing close guard applies unchanged.
- fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold
identity fits; the slug-shape security property is unchanged.
- tests: cross-origin closure through the real runner seam, an 81-char
identity through the adapter, cap and shape refusals, routed-work skips,
nonexistent-identity skips, and idempotent replay.
* feat(bearings): add the /bearings lavish interactive fleet board
/bearings lavish renders the bearings snapshot onto a shipped, reusable board
template and arms it as a Lavish process-event source, so the captain answers
Captain's Call items on the board and firstmate is woken by an ordinary check
wake - no conversational turn ever blocks on a poll.
- .agents/skills/bearings/assets/board-template.html: the shipped template
(myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot,
fail-closed schema guard that renders an error card instead of an empty
fleet). Per-invocation agent work is composing the payload only.
- bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload
validation, slot injection with a round-trip check and \u003c escaping,
stable board path, any-origin bind ALWAYS before arm, arm-if-absent.
- bearings SKILL.md: the lavish invocation option, board composition rules,
board-wake handling, and the captain-ruled merge-click authorization with
its mandatory safeguards (PR resolved from the task's own meta record,
wake-time green re-verification, never a red or changed PR, merges only
through bin/fm-pr-merge.sh, chat echo with the full PR URL).
- process-event-sources SKILL.md: one-line board-wake routing trigger.
- tests: payload refusals, injection round-trip, bind-before-arm, idempotent
re-arm, and template slot integrity.
Fleet pickup: homes receive this after merge plus a firstmate self-update;
landing timing is coordinated with the main firstmate.
* no-mistakes(review): Harden bearings board validation and wake handling
* no-mistakes(review): Require HTTPS for bearings board PR links
* no-mistakes(review): Fail closed and bound bearings board answers
* no-mistakes(review): Enforce UTF-8 byte limits for board answers
* no-mistakes(review): Serve bearings board before arming and reject empty actions
* no-mistakes(review): Prove bind-before-arm ordering through live answer consumption
* no-mistakes(document): Document bearings board and cross-origin answers
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
---------
Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
Co-authored-by: Mickaël Rémond <mremond@process-one.net>
Co-authored-by: Inthuson <iaminthuson@gmail.com>
Co-authored-by: Inthuson <inthuson@amazon.com>
Bre77
added a commit
to Bre77/firstmate
that referenced
this pull request
Aug 22, 2026
…s) (#70) * fix(ci): fail hung Herdr behavior runs in 20 minutes (#2413) A wedged family-run step was occupying the runner until the 75-minute job cap; bound that step so cleanup and timing artifacts still upload. * fix: keep the public promise reachable when work is routed to a second mate (#2457) The lightweight Relay follow-up link lives in the answering home's own state/<task-id>.meta, so it can only bind work that home owns. When a Relay-linked request is routed to a second mate, the task record lives in the second mate's home, fm-x-link.sh failed with a bare "no such task ...meta", and nothing else picked the promise up: only the soft acknowledgement was ever posted. The typed promised-final path already supports --work-home secondmate:<id>; the playbook simply never chose it. - fmx-respond now states the routing rule crisply: a task in this home takes the lightweight link, and second-mate-routed work takes a promised-final commitment bound to that home, registered up front with the brief command carried into the routed worker's instructions. - fm-x-link.sh refuses a task with no local record by naming the registered second mate whose home actually holds it and printing the promised-final registration command, with the exact --work-home when the match is unambiguous. A home with no registered second mates keeps the plain error. - fm-backlog-handoff.sh reports, after a successful move, any moved key that still owes a public reply bound to main/<key>, since that binding no longer names the home owning the work. The move itself is never blocked. Docs and the secondmate handoff prose follow the same rule. Tests cover the refusal, its scoping, the unchanged local-link path, and both handoff outcomes at the script boundary. * docs(skills): add remote-secondmate recovery hint for false-negative verdicts (#2456) * fix(skills): hint that remote secondmate liveness verdicts false-negative fm-crew-state and fm-send routinely misreport a live remote secondmate as dead; confirm against the pane before relaunching, and relaunch only through fm-spawn.sh, never raw herdr pane surgery. * no-mistakes: apply CI fixes * fix(calm): keep Pi's export confirmation visible (#2461) Pi 0.83.0 added a status line to every tool-expansion change, and Pi updates the previous status line in place when two status messages arrive back to back. Calm's post-export redraw cycled tool expansion on the macrotask right after Pi printed "Session exported to: <path>", so both expansion status lines coalesced over that confirmation and the captain was left with no record of where their export landed. Calm now repaints only the tool rows it presents, by invalidating each row through the render context Pi hands its render slots, and requests the surrounding redraw through setStatus. Neither appends to the transcript. The repaint is still needed because Pi can re-render a row asynchronously - the built-in edit row invalidates itself once its diff is ready - and that re-render can land inside the window where /export forces stock rendering. The real-terminal /export case now asserts the confirmation is still on screen after the redraw has settled, and that the redraw restored every Calm-hidden row, instead of only racing the moment the confirmation first appeared. * feat(stow): add open-record persistence to /stow before reset (#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (#2515) * ci: gate GitHub workflows with pinned actionlint (#2517) * fix(lint): catch malformed GitHub workflows before merge A self-broken ci.yml cannot report its own breakage, so parse every workflow in the local lint path that no-mistakes already runs. * fix(lint): pin actionlint instead of Ruby for workflow lint A self-broken ci.yml still has to fail in the local lint path, and the named tool for that gate is actionlint, not a new Ruby runtime. * no-mistakes(document): Clarify pinned workflow lint documentation * fix: install pinned lint tools across supported platforms (#2546) * fix: install pinned shellcheck and actionlint on macOS and linux arm64 The installers were hardcoded to linux amd64 and sha256sum, so a Mac dev could not satisfy the refuse-on-mismatch lint gate. Select the official per-platform archive and checksum, and fall back to shasum -a 256. * no-mistakes(document): Document cross-platform pinned lint installers * docs: reconcile test-evidence docs with store_in_repo: true (#2548) .no-mistakes.yaml has set test.evidence.store_in_repo: true since #2355, but CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described the old policy of keeping evidence out of the repo in a temp directory. The current no-mistakes behavior for store_in_repo: true is to publish each run's test evidence to the orphan no-mistakes/evidence branch and link it from the PR body. That branch shares no history with code branches, so evidence never enters a pushed feature branch or the default branch, and CI's tracked personal fleet paths rule stays accurate. Docs only. No change to .no-mistakes.yaml or any workflow. * docs: clarify test evidence branch storage (#2549) * docs: correct test evidence storage comment in .no-mistakes.yaml * no-mistakes: apply CI fixes * docs: hint that live scouts may host their own Lavish review loop (#2563) Make that a first-class option in always-loaded instructions so firstmate does not default to mediating and tearing the scout down between iteration rounds. * fix(bin): report remote secondmate delivery and state truthfully (#2570) * fix(bin): report remote secondmate delivery and state truthfully A steer to a remote secondmate crosses fm-on.sh to a host-local fm-send leg whose unconfirmed submit read-back (verdict=pending, typically a busy mate whose harness queues the steer) was flattened into exit 1, so the parent printed "error: text not submitted" / "error: text not sent" and discarded the pending-reply expectation for a steer that had actually landed. fm-send now carries the verdict across the ssh boundary as a documented delivered-unconfirmed exit 3: the parent reports the steer as delivered with confirmation pending, exits 0, keeps the expectation armed (awaiting_report), and closes --resolve-key decisions, while transport loss (ssh 255) and real remote failures keep failing loudly with the remote leg's stderr attached. A local unconfirmed submit now also exits 3 with an honest non-error message and still never closes a decision key. fm-crew-state.sh and fm-peek.sh no longer read a remote mate's endpoint through local probes (which misreported a healthy mate as "worktree gone" / "can't find session: remote"): both now use the true remote source over fm-on.sh, and an unreachable or unreadable remote reads as unknown-remote, never as gone or dead. * no-mistakes(document): Document remote delivery and state truth * no-mistakes: apply CI fixes * feat: adopt spendPriority for quota dispatch (#2574) * Adopt quota-axi 0.1.29 spendPriority-primary array dispatch. quota-axi 0.1.29 publishes schema 5 with selection.spendPriority as the primary comparative signal and demotes derivation fields out of default --json. Rank comparable-fit candidates on that scalar, keep runway versus the completion horizon as a hard gate, and raise the compatibility floor so a pre-consolidation build cannot reach dispatch intake. * no-mistakes(review): Correct schema fixtures and remove prescriptive selection prompts * no-mistakes(document): Correct quota verification evidence chronology * Collapse quota-array-dispatch onto TOON-first spendPriority ranking. Decide from quota-axi's default TOON; keep --json as a rare defensive fallback. Rank by spendPriority after eligibility, reasoning-class, and runway-feasibility gates, and drop the hand-computed Pareto, pace, reserve, and window-id layers. * no-mistakes(review): Permit ambiguous JSON fallback and correct reset fixtures * no-mistakes(review): Correct runway semantics and escalate unresolved uncertainty * no-mistakes(document): Document TOON-first quota dispatch evidence * docs: add GROK_BOT.md Grok Bot system prompt (#2590) * docs: add GROK_BOT.md Grok Bot system prompt * docs: amend GROK_BOT.md with charter report-back and delegation marker * docs: classify GROK_BOT.md as public-product * docs: make GROK_BOT.md the plain Grok Bot system prompt * docs: update GROK_BOT.md nautical terms and self-improvement (#2592) * doc: Update language in GROK_BOT.md for clarity Refine language for clarity and consistency in instructions. * fix(bin): preserve inactive reconciliation scan progress (#2595) * fix(bin): guarantee inactive-reconcile scan progress under second quantization The inactive-outcome scan computed its aggregate deadline in whole seconds, so a 1-second budget's effective value lands anywhere in (0,1]; a scan starting just before a wall-clock second boundary rounded its whole budget away mid-scan and exited having visited no child, while the durable cursor had already advanced past the never-examined child. This is the CI flake behind tests/fm-inactive-reconcile.test.sh's 'next bounded scan did not resume with the following child' (watcher-wake-lock family, portable serial 2, seen on the PR #2590 run). Every scan now visits at least its first due child with the per-child state-read bound floored at one second, so no invocation can be a zero-work no-op. The outer process-group kill moves to budget+1s: the scan's own deadline enforces the budget, and the kill is a backstop for a scan wedged in an unbounded wait instead of a racer that routinely preempts the clean bounded exit. The wake-lock-wait test bound tracks the backstop (3s -> 4s); the previously flaky assertion is unchanged. * no-mistakes(document): Document inactive-reconcile deadline backstop * doc: Revise Firstmate delegation and communication guidelines Refactor the guidelines for Firstmate's role and delegation process, emphasizing the importance of crewmates and asynchronous work. * doc: Update work delegation and secret management instructions Clarified guidelines for handing off work to crewmates and managing secrets. * test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617) Three assertions in tests/fm-procevent.test.sh depended on a detached runner having finished work that the command starting it does not wait for. reconcile's replacement runner is started through detach_runner, which only forks: reconcile returns and counts the start before that runner has claimed its source or exec'd its child. Any assertion taken straight after reconcile therefore samples a race. - The publish-before-apply recovery section left its always-ready /bin/echo source registered across the recovery reconcile, so that reconcile launched a competing detached poll (observed: started=1) that then raced every later assertion for the source claim, the next capture sequence, and this home's applied record, and outlived the section holding a live claim. It is now retired before that reconcile - re-announcement is proven from the durable inbox alone and needs no registration - and started=0 is asserted so a competing poll cannot be reintroduced unnoticed. This is the same retire-before-reconcile discipline the self-announcing section already carries; that section acquired it after the identical race made its "not-autohandled: self-src" assertion read "already owned: self-src". - The crashed-leader replacement section snapshotted the replacement's claim file and execution log behind a fixed 0.5s settle window. On a loaded machine that window expires first, which is the CI flake behind "a replacement runner started without recording its own claim" and "reconcile did not start exactly one replacement source". Both effects are now waited for with the suite's bounded wait helpers; the exact one-replacement count is still asserted afterwards, unchanged. - The duplicate-start section slept 0.5s for reconcile's runner to record ownership before asserting that a second start loses to it. It now waits for that claim. Also tighten one assertion that could not fail as written: "autohandled: self-src" is a substring of "not-autohandled: self-src", so the applied path was accepted even when the runner reported the capture left for the handler. Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced 6 failing runs, all in the crashed-leader section. On the fixed suite, 216 full runs under the same load produced none. Reverting the self-announcing section's retire-before-reconcile line reproduces "already owned: self-src" on the first iteration, confirming the shared mechanism. * fix: preserve pending replies and defer remote reposts (#2618) * fix(bin): keep pending-reply expectations honest on both send legs Two related asymmetries let the parent-owned secondmate reply guard drop or nag requests it should not have. Local delivered-unconfirmed dropped the expectation. A marked request whose submit read-back stayed unconfirmed (verdict=pending) is the same not-a-failure outcome the remote leg reports as delivered, but fm-send discarded the parent's pending-reply record for it, so a request that very likely landed stopped being tracked entirely. The record now stays armed on its unconfirmed-delivery marker: a correlated report still resolves it, and an unanswered one still surfaces through the library's own reconciliation. Exit 3 and the local rule that an unconfirmed answer never closes a decision key are unchanged. Remote replies were nagged for a repost they did not need. A remote mate's report reaches the parent's status log only through the asynchronous mirror in fm-procevent-remote-reply.sh, yet the guard read an absent correlated line as proof the mate never reported - even while the answer was still in flight, which is the common case because the mirror's poll window is comparable to the recovery grace. The mirror now publishes one caught-up watermark from a quiet window, and the guard admits a missing report as evidence only once that watermark passes the turn that should have produced it. A genuinely missed report still gets exactly one repost, and a channel that is behind, unarmed, or broken leaves the request durably open and un-nagged rather than nagging blind; the mirror escalates its own continuity failures as before. Tests: a local unconfirmed secondmate send keeps its expectation armed and resolvable; a mirrored correlated remote reply resolves with no repost; a stale or absent watermark withholds the repost while a fresh one still releases it; a quiet remote window publishes the watermark and retirement clears it. * no-mistakes(review): Distinguish preempted polls from quiet windows * no-mistakes(document): Clarify remote reply channel freshness * no-mistakes(lint): Annotate shared remote preemption exit constant * fix(bin): honor declared pauses in busy-pane wedge checks (#2619) * fix(watch): honor a declared pause on a busy pane's completed-turn bound A worker that declares an external wait (`paused:`) and then blocks in one long foreground call - a review-hosting scout parked in a single blocking `lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane BUSY, so the stale path that already honors declared pauses never ran for it. The busy-pane completed-turn bound instead routed it straight into wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long as the review stayed open. busy_turn_bound_check now owns which absorber takes a crossed bound: a crew whose own last status line declares an external wait or a verified captain-held transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else keeps the unchanged wedge timer. The discriminator is the declaration together with liveness (the caller has already confirmed the pane is busy), never a blanket silencing - a crew that declared nothing, or whose pane is not live, escalates exactly as before, and a declared pause still re-surfaces once per long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched: the daemon owns pause triage there and already reads the same vocabulary. The two call sites also no longer clear pause bookkeeping in the same poll the pause cadence recorded it, which would have erased the re-surface throttle and turned the long cadence back into a per-poll re-surface. Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence recheck, and the restored wedge escalation once the declaration is lifted on the same busy over-age pane. Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded machine: fixed liveness budgets were reaping watchers mid-startup, so assertions on post-poll state passed vacuously or failed spuriously. Waits that describe a poll's outcome now wait for a completed poll cycle via the liveness beacon, the heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit budget is the uniform 10s already used elsewhere in the file. * no-mistakes(review): Fail poll-cycle waits on timeout * no-mistakes(review): Prevent poll timeout test hangs * no-mistakes(document): Clarify paused busy-pane supervision * doc: Enhance communication guidelines for decision-making Added guidelines for decision communication to the captain. * doc: Update task delegation and communication guidelines Clarify communication protocols with crewmates regarding task delegation and reporting. * fix(bin): reliably confirm herdr steer submission (#2647) * fix(herdr): confirm local steers that native agent-state misses Herdr can leave agent_status idle for a landed Claude turn and can keep queued Enter text visible while busy, so fm-send was reporting false swallows. Confirm those cases through the shared queued-Enter verdict and a cleared composer, and keep a genuine idle pending composer as unconfirmed. * no-mistakes(review): Stop Herdr Enter retries on unreadable composers * no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail * no-mistakes(review): Prevent confirmation after failed Herdr Enter * no-mistakes(review): Pace Herdr retries and clarify submit fallback * no-mistakes(review): Align Herdr submit docs with idle fallback * no-mistakes(document): Correct Herdr submit-confirmation documentation * feat(bearings): add interactive Lavish fleet board (#2659) * feat(bin): accept any-origin decision bindings with full-identity keys An aggregation surface (the bearings board) carries captain answers for holds across origins, but a binding was one-origin-per-source and the Lavish adapter capped question keys at 64 chars while real full hold identities measure 69-81. - fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker; binding prints it verbatim and answers accepts it, so the runner's feed seam carries an any-origin source with no runner change. In any-origin mode each key is a full hold identity <origin>-decision-<key>, split at its first -decision-; a key with no separator (merge/dispatch instructions) is skipped and feeds nothing, keeping non-decision answers out of the hold ledger by construction. Every existing close guard applies unchanged. - fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold identity fits; the slug-shape security property is unchanged. - tests: cross-origin closure through the real runner seam, an 81-char identity through the adapter, cap and shape refusals, routed-work skips, nonexistent-identity skips, and idempotent replay. * feat(bearings): add the /bearings lavish interactive fleet board /bearings lavish renders the bearings snapshot onto a shipped, reusable board template and arms it as a Lavish process-event source, so the captain answers Captain's Call items on the board and firstmate is woken by an ordinary check wake - no conversational turn ever blocks on a poll. - .agents/skills/bearings/assets/board-template.html: the shipped template (myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot, fail-closed schema guard that renders an error card instead of an empty fleet). Per-invocation agent work is composing the payload only. - bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload validation, slot injection with a round-trip check and \u003c escaping, stable board path, any-origin bind ALWAYS before arm, arm-if-absent. - bearings SKILL.md: the lavish invocation option, board composition rules, board-wake handling, and the captain-ruled merge-click authorization with its mandatory safeguards (PR resolved from the task's own meta record, wake-time green re-verification, never a red or changed PR, merges only through bin/fm-pr-merge.sh, chat echo with the full PR URL). - process-event-sources SKILL.md: one-line board-wake routing trigger. - tests: payload refusals, injection round-trip, bind-before-arm, idempotent re-arm, and template slot integrity. Fleet pickup: homes receive this after merge plus a firstmate self-update; landing timing is coordinated with the main firstmate. * no-mistakes(review): Harden bearings board validation and wake handling * no-mistakes(review): Require HTTPS for bearings board PR links * no-mistakes(review): Fail closed and bound bearings board answers * no-mistakes(review): Enforce UTF-8 byte limits for board answers * no-mistakes(review): Serve bearings board before arming and reject empty actions * no-mistakes(review): Prove bind-before-arm ordering through live answer consumption * no-mistakes(document): Document bearings board and cross-origin answers * fix(bearings): restore decision options and add close controls (#2707) * fix(bearings): always show decision options and a close/drop control Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call. * no-mistakes(review): Fix drop closure and decision-only option validation * no-mistakes(review): Preserve answerability for non-decision cards * no-mistakes(document): Clarify decision drop documentation * ci: require no-mistakes pipeline step attestation (#2710) Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed. * feat: collapse decisions into tasks held for the captain (#2728) * feat(captain-hold): collapse the decisions concept into tasks held for the captain A decision is no longer a separate type: it is an ordinary backlog task held for the captain, identified by its task id. bin/fm-captain-hold.sh owns the surviving behaviors - guarded hold creation, the recorded-answer close (answer/answers with a release mode for captain-gated work), the source bindings, and the investigation completion gate - and bin/fm-decision-hold.sh becomes a one-release compatibility shim over it. The fleet snapshot now parses hold-until and computes captain_actionable as queued + captain-held + unblocked + due, independent of row kind, plus a presentation-only deferred_marker for prose-deferred rows. Bearings renders every due captain-held task in Captain's Call, date-deferred holds as dated Charted Next gates, suppresses prose-deferred rows from default views with an omitted disclosure, and excludes from Recently Landed anything that closed while still held for the captain. Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already plain task ids and keep working; short keys in recorded metadata, concrete origin bindings, chat --resolve-key fallbacks, and old resolution records all resolve in place. * no-mistakes(review): Fix captain answer replay and body preservation * no-mistakes(review): Fix captain hold idempotency and legacy replay * no-mistakes(review): Validate card close modes and compatibility routing * no-mistakes(review): Enforce release replay mode matching * no-mistakes(review): Prevent duplicate decision cards and released replay mismatches * no-mistakes(review): Preserve answer columns and legacy resolve replays * no-mistakes(document): Document strict replay and legacy compatibility * no-mistakes(lint): Quote done literals to satisfy ShellCheck * no-mistakes: apply CI fixes * fix(rebase): keep collapsed captain hold board semantics * fix: bound recovery announcements and preserve supervision (#2733) * fix(watch): announce recovery once per generation and keep successors supervising A lost Pi/OpenCode handling handshake re-announced the same recovery generation on every cycle and spent the successor's first ~55s blind, so a real crew event could be ignored and then dropped. Record the announcement in the durable marker, confirm the handshake before the follow-up without swallowing failure, and enter the poll loop immediately. * no-mistakes(review): Tighten recovery event timing regression * no-mistakes(document): Document recovery-loop supervision guarantees * fix(bin): surface captain-call record divergence (#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue #2251's PR #2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after…
doitdigital0495
added a commit
to doitdigital0495/firstmate
that referenced
this pull request
Aug 23, 2026
* fix(bearings): restore decision options and add close controls (kunchenguid#2707) * fix(bearings): always show decision options and a close/drop control Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call. * no-mistakes(review): Fix drop closure and decision-only option validation * no-mistakes(review): Preserve answerability for non-decision cards * no-mistakes(document): Clarify decision drop documentation * ci: require no-mistakes pipeline step attestation (kunchenguid#2710) Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed. * feat: collapse decisions into tasks held for the captain (kunchenguid#2728) * feat(captain-hold): collapse the decisions concept into tasks held for the captain A decision is no longer a separate type: it is an ordinary backlog task held for the captain, identified by its task id. bin/fm-captain-hold.sh owns the surviving behaviors - guarded hold creation, the recorded-answer close (answer/answers with a release mode for captain-gated work), the source bindings, and the investigation completion gate - and bin/fm-decision-hold.sh becomes a one-release compatibility shim over it. The fleet snapshot now parses hold-until and computes captain_actionable as queued + captain-held + unblocked + due, independent of row kind, plus a presentation-only deferred_marker for prose-deferred rows. Bearings renders every due captain-held task in Captain's Call, date-deferred holds as dated Charted Next gates, suppresses prose-deferred rows from default views with an omitted disclosure, and excludes from Recently Landed anything that closed while still held for the captain. Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already plain task ids and keep working; short keys in recorded metadata, concrete origin bindings, chat --resolve-key fallbacks, and old resolution records all resolve in place. * no-mistakes(review): Fix captain answer replay and body preservation * no-mistakes(review): Fix captain hold idempotency and legacy replay * no-mistakes(review): Validate card close modes and compatibility routing * no-mistakes(review): Enforce release replay mode matching * no-mistakes(review): Prevent duplicate decision cards and released replay mismatches * no-mistakes(review): Preserve answer columns and legacy resolve replays * no-mistakes(document): Document strict replay and legacy compatibility * no-mistakes(lint): Quote done literals to satisfy ShellCheck * no-mistakes: apply CI fixes * fix(rebase): keep collapsed captain hold board semantics * fix: bound recovery announcements and preserve supervision (kunchenguid#2733) * fix(watch): announce recovery once per generation and keep successors supervising A lost Pi/OpenCode handling handshake re-announced the same recovery generation on every cycle and spent the successor's first ~55s blind, so a real crew event could be ignored and then dropped. Record the announcement in the durable marker, confirm the handshake before the follow-up without swallowing failure, and enter the poll loop immediately. * no-mistakes(review): Tighten recovery event timing regression * no-mistakes(document): Document recovery-loop supervision guarantees * fix(bin): surface captain-call record divergence (kunchenguid#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (kunchenguid#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue kunchenguid#2251's PR kunchenguid#2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (kunchenguid#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (kunchenguid#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (kunchenguid#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (kunchenguid#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (kunchenguid#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (kunchenguid#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (kunchenguid#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (kunchenguid#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in kunchenguid#2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e79 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after kunchenguid#2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (kunchenguid#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (kunchenguid#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (kunchenguid#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (kunchenguid#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (kunchenguid#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (kunchenguid#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (kunchenguid#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (kunchenguid#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * no-mistakes(document): fix portable serial shard counts after upstream merge * no-mistakes: apply CI fixes --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Mickaël Rémond <mremond@process-one.net> Co-authored-by: Inthuson <iaminthuson@gmail.com> Co-authored-by: Inthuson <inthuson@amazon.com> Co-authored-by: Christopher McKay <101884182+karotkriss@users.noreply.github.com> Co-authored-by: Daan Aerts <daan@daeverhuur.nl>
Marl0nL
pushed a commit
to Marl0nL/firstmate
that referenced
this pull request
Aug 24, 2026
…he merge call Phase 5 salvage of the second half of fork fix b810a37 (#42). Upstream 5b6d0fb (kunchenguid#2779) independently closed the -R/-dR bundled-cluster half of #42 (already on the trunk), but the trunk's reject_repo_overrides lets --hostname pass straight through its --*) arm, and the github merge call sets no GH_HOST. --hostname is a gh-axi/gh global flag that sets GH_HOST for the child gh, so an extra --hostname arg merges the recorded PR against a different GitHub host while state still records the github.com URL - a merge-authority host redirect the trunk did not guard. Add --hostname|--hostname=* to reject_repo_overrides and pin GH_HOST=github.com on the github merge call. Correct the cluster-guard comment's rationale (gh-axi does not currently expand/forward such a cluster; the guard stays correct defensively either way), which #42's own review flagged. New tests: --hostname and --hostname=<value> refuse before recording or reading state, and the successful github merge is pinned to GH_HOST=github.com (both mutation-verified). The GitLab path already scopes glab by the URL host.
Marl0nL
added a commit
to Marl0nL/firstmate
that referenced
this pull request
Aug 24, 2026
…nk (#47) * test(bootstrap): isolate the orca-backend gate case from an ambient orca binary Phase 5 salvage of fork fix 9a8ecfd. The orca-backend-selected case builds its PATH from $BASE_PATH (which includes /usr/bin), so on a host that has an unrelated `orca` program installed (GNOME's screen reader lives at /usr/bin/orca) bootstrap sees orca present and does not emit the expected "MISSING: orca" line, failing the assertion spuriously. Upstream never fixed this. Add make_orca_free_sysbin so the case observes orca as genuinely absent, proving the gate requires it regardless of the ambient system. Reproduced failing on the integration trunk before this change. * fix(brief): forbid agent commit co-author in ship, scout, and charter scaffolds Phase 5 salvage of fork fix ff2c958. Project and secondmate crews never load firstmate-coding-guidelines (a firstmate-repo-only trigger), so the fork's no-agent-co-author rule is invisible to them and every commit falls back to the harness default Co-Authored-By trailer. Upstream never pushed the rule into briefs. State it directly in the generated ship, scout, and secondmate charter scaffolds, where commit-capable crews actually read it. Note for review: this encodes the fork's standing policy (no agent co-author on ANY crew commit, AGENTS.md), which is broader than upstream's firstmate-repo-only ban. * fix(secondmates): truncate the registry charter-summary to a scannable one-liner Phase 5 salvage of fork fix 8b25fc8, re-homed onto upstream's bin/fm-secondmate-charter-lib.sh (the logic moved there from fm-home-seed.sh). data/secondmates.md prints in full at every session start in every home with a secondmate, so a multi-thousand-word charter inlined as the summary field burns context forever. Deterministically truncate the summary to a short one-liner (REGISTRY_SUMMARY_CAP=200, word-boundary cut, no model call); a caller can pass FM_SECONDMATE_SUMMARY directly. The full charter is still copied verbatim to the home's data/charter.md, so nothing is lost. The scope field, which the main firstmate uses for routing, is left untouched. The parenthesis/semicolon sanitizer stays, documented as protecting the registry field-split parser. Upstream never truncated the summary. New colocated lib test covers the cap, short passthrough, the explicit override, and that scope is never truncated. * fix(herdr-lab): scope lab calls with a leading --session so agent start cannot leak into default Phase 5 salvage of fork fix ff80250. bin/fm-herdr-lab.sh promises a named non-default isolated session for every Herdr lifecycle action (AGENTS.md section 11 treats it as a safety contract), but fm_herdr_lab_raw appended --session <lab> at the END of the argument list. For `agent start <name> ... -- <argv>` the flag therefore landed after the `--` separator, inside the agent's own argv, leaving the Herdr call itself unscoped and creating the agent in the ambient (live default) session. Compose every call as `herdr --session <lab> <args...>` with the flag leading, the position Herdr's own usage documents: it is consumed as a global option before subcommand dispatch, so no positional operand and nothing after a `--` separator can capture it. Also fail closed: fm_herdr_lab_raw re-validates the lab name and refuses any argument list carrying its own --session anywhere. Upstream's fm-herdr-lab.sh was byte-identical to the pre-fix fork state (still trailing). The adapter path bin/backends/herdr.sh keeps its trailing --session and is unaffected: it forwards no `--` command and never calls agent start. Regression coverage flips the fake-herdr to resolve --session as a global flag stopping at `--`, and adds the agent-start leak/target assertions. * fix(spawn): refuse a worktree another task of this home already owns Phase 5 salvage of fork fix fea2afe. Twice on 2026-07-29 a spawn was handed a worktree a LIVE task was still working in: in the main home a herdr-server restart left one task's meta pointing at another task's worktree (the confused crewmate wrote its uncommitted fixes there, where they were reverted as foreign edits and lost); in a secondmate home a spawn re-leased a live crewmate's worktree and `treehouse get` reset it to the default branch. The pool judges a slot free from live process cwds, so a crashed-but-owned slot (or an agent whose cwd is the pane's launch dir) looks free. This home's own state/*.meta is the record that knows the slot is taken. fm-spawn now refuses to launch into a worktree ANY other task of the current FM_HOME records as `worktree=`, naming the owning task. A respawn of the SAME task id into its own recorded worktree stays legitimate. The check runs inside the isolation assertion (both tmux and Orca ship/scout paths) and, on the secondmate path, before the pre-launch fast-forward and config push. It is deliberately home-scoped and never sweeps sibling homes. Upstream never added this guard (the closest, kunchenguid#765 two-stable-reads, settles same-task path races, not cross-task ownership). New colocated lease test; the dispatch-profile batch and path-spelling fixtures are corrected to give each live task its own slot, since two live tasks must never share one. * fix(paths): respect treehouse's path spelling, canonicalize our own Phase 5 salvage of fork fix 8107742. Two path-aliasing bugs with one root cause and opposite fixes. Where /home is a symlink to /var/home (the default on every ostree/atomic Fedora variant), one directory has two spellings, and whether to canonicalize depends on who owns the spelling. teardown: `treehouse return` string-matches its argument against the spelling treehouse recorded, so every teardown failed on such a box with "is not managed by treehouse". Firstmate never receives that spelling for a crew worktree (the worktree is found by polling the pane's cwd, which every backend reports OS-resolved), so ask treehouse's own inventory which spelling is its, matching on physical identity, at the handoff boundary - which also repairs metas written before this fix. Degrades to the caller's path when treehouse cannot be asked (fm-teardown.sh: treehouse_recorded_path). turn-end guard: fm_watcher_lock_matches_pid string-compared two of firstmate's OWN values, so a watcher armed from one spelling and a hook invoked through the other failed to match and the guard blocked turn ends at a live, beating watcher. Both sides are ours, so canonicalize both before comparing (fm-wake-lib.sh: fm_same_path; same fix in fm-watch-arm.sh). The matcher stays strict: a genuinely different home still fails. Upstream never handled the /home->/var/home aliasing (the closest, kunchenguid#765, settles same-task path-read races). Recorded in docs/treehouse-path-contract.md. Tests build the alias with ln -s so they hold on a box whose /home is not a symlink. * fix(backlog): name this home's backlog explicitly in hand-run tasks-axi guidance Phase 5 narrow salvage of fork fix 1b9fbf6. tasks-axi resolves both its .tasks.toml and the backlog path from the process working directory, so a tasks-axi call hand-run from a shell standing inside a project clone reads that clone's absent backlog as empty (a confident wrong answer) and can write a stray backlog.md into the clone - a forbidden project write. Upstream already pins every AUTOMATED bin/ call site to an explicit backlog file (fm-backlog-handoff.sh --file, the captain/decision-hold cd-into-home wrappers, session-start --file), so the fork's shared fm_backlog_file/fm_tasks_axi_run library rewrite is NOT re-applied - it would duplicate a solution upstream reached differently. Salvage only the residual instruction surface upstream lacks: the AGENTS.md section 10 rule that every hand-run tasks-axi call names `--file "$FM_HOME/data/backlog.md"`, and fm-teardown's post-teardown backlog reminder, whose printed done/ready commands a firstmate copy-runs, now naming the home backlog explicitly. * fix(supervision): stop false-wedging a finished crew that appends resolved: after done:, and stop churning stale wakes for a crew parked awaiting merge Phase 5 salvage of fork fixes c6b215c (#10) and c2c9168 (#12), bundled because #12 builds on #10's real-state reading. #10: AGENTS.md section 11 tells a crew that opened a keyed decision or blocker to durably close it with a `resolved:` line, so a crew that finished correctly ends its log `done: ...` then `resolved: ...`. The terminal-state scan read the raw last line, so it misread that finished log as non-terminal and false-wedged an idle-but-done pane. Add status_line_is_resolved + last_state_status_line to the one shared owner (fm-classify-lib.sh) and route stale_is_terminal, the away daemon's classify_stale / mark_escalated_seen / handle_wake, and crew-state's resolved-line promotion through it. `done:` then `resolved:` reads terminal; `working:` then `resolved:` (a mid-task blocker close) stays non-terminal. #12: a done crew with an armed merge-monitor (state/<id>.check.sh) idles until its PR merges; its redraw-jittered idle pane is not byte-stable, so the stale seam would re-surface it as a fresh possible-wedge on every distinct hash. Add crew_is_parked_awaiting_merge (reconciled done + armed check, re-read every evaluation, fail-closed) and consult it as the first absorb in both the always-on watcher's terminal-stale branch and the daemon's classify_stale, absorbing with no wedge timer. A re-activated crew (busy pane, or its verb moved off done) fails the predicate and full stale sensitivity resumes with no manual re-arm. Upstream's Stop-auto-arm continuity model did not close either gap (its classify_stale still falls a done:+resolved: log through to transient-stale). * fix(spawn): fail loudly when a spawn creates a pane but no agent starts Phase 5 salvage of fork fix 723ab78. fm-spawn printed `spawned` as soon as it typed the launch command; a launch send that never landed (observed live against a freshly restarted herdr server) left a bare no-agent shell yet still reported success and went In flight - a silent failure, especially dangerous in away mode. After sending the launch command, poll fm_backend_agent_alive until the agent is confirmed running before printing `spawned`. A confidently dead pane triggers one launch re-send (race mitigation) and, if still dead, a LOUD non-zero failure that names the pane. The check runs only for a backend+harness whose liveness probe can reach a confident verdict (new fm_backend_agent_probe_verifiable: herdr for any harness, tmux for non-pi harnesses); an inconclusive `unknown` proceeds with a warning (never a false failure), and unverifiable backends keep the pre-confirmation behaviour (a documented gap). Kimi already confirms its own ready+delivery, so it is skipped. Upstream added no general spawn confirmation (only a kimi-specific one). FM_SPAWN_CONFIRM=0 disables it; tests/lib.sh defaults it off so every fake-pane spawn test skips the poll. * test(spawn): release the pool-base lease before the idempotent-repeat spawn Fixture correction belonging with the worktree-lease guard (fdc9099): the idempotent-repeat case re-spawns a second task id into the same pooled base to prove the base refresh is idempotent, which the new lease guard correctly refuses (two live tasks must never record the same worktree). Model the first task's teardown by removing its meta before the repeat, exactly as the dispatch-profile path-spelling fixture does. Found by the full-suite run; the test passes on the base and failed only under the guard. * fix(fm-pr-merge): guard --hostname host-redirect and pin GH_HOST on the merge call Phase 5 salvage of the second half of fork fix b810a37 (#42). Upstream 5b6d0fb (kunchenguid#2779) independently closed the -R/-dR bundled-cluster half of #42 (already on the trunk), but the trunk's reject_repo_overrides lets --hostname pass straight through its --*) arm, and the github merge call sets no GH_HOST. --hostname is a gh-axi/gh global flag that sets GH_HOST for the child gh, so an extra --hostname arg merges the recorded PR against a different GitHub host while state still records the github.com URL - a merge-authority host redirect the trunk did not guard. Add --hostname|--hostname=* to reject_repo_overrides and pin GH_HOST=github.com on the github merge call. Correct the cluster-guard comment's rationale (gh-axi does not currently expand/forward such a cluster; the guard stays correct defensively either way), which #42's own review flagged. New tests: --hostname and --hostname=<value> refuse before recording or reading state, and the successful github merge is pinned to GH_HOST=github.com (both mutation-verified). The GitLab path already scopes glab by the URL host. * docs/style: one sentence per line, trim duplicated contracts, tidy a log string Review cleanups (non-blocking findings 4-6, 8): - docs/treehouse-path-contract.md: split the three multi-sentence audit bullets to one sentence per line (repo style). - One-owner: the treehouse path contract is now stated in full only in docs/treehouse-path-contract.md; the fm_same_path (fm-wake-lib.sh) and treehouse_recorded_path (fm-teardown.sh) headers keep the load-bearing "do NOT canonicalize a path an external tool owns - that is the bug" note plus a one-line why and a cross-ref, dropping the restated ostree-aliasing story. - One-owner: the "scope is never truncated because it is the routing input" rationale lives once in docs/configuration.md; fm-home-seed.sh's header states the fact and points there. - fm-supervise-daemon.sh: classify_stale's transient-stale log line now prints $term (the real-state line the decision is made on), not the raw $last, which could show a trailing resolved: bookkeeping line. No behaviour change. --------- Co-authored-by: Marlon <marlonleicester@gmail.com>
IanQiu979
added a commit
to IanQiu979/firstmate
that referenced
this pull request
Aug 25, 2026
…'s landed-work check (#2) * feat(bin): merge GitLab merge requests through the guarded PR merge path (kunchenguid#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (kunchenguid#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in kunchenguid#2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e79 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after kunchenguid#2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (kunchenguid#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (kunchenguid#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (kunchenguid#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (kunchenguid#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (kunchenguid#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (kunchenguid#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (kunchenguid#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (kunchenguid#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (kunchenguid#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (kunchenguid#2857) * feat(bin): deliver local steers through durable task inboxes (kunchenguid#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (kunchenguid#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (kunchenguid#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: kunchenguid#2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (kunchenguid#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (kunchenguid#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (kunchenguid#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes kunchenguid#2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (kunchenguid#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (kunchenguid#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (kunchenguid#2939) * fix(bin): bound remote job worker supervisor restarts (kunchenguid#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix(teardown): recognize work that landed under rewritten commits fm-teardown.sh proved landed work by asking whether the branch's own commits are reachable from the default branch. A rebase gives every commit a new object id, so once landing rewrote them that question can never be answered yes again, and teardown refused work that was fully in main. Because bin/fm-merge-local.sh is fast-forward-only, rebasing onto a moved default branch is the normal way a local-only chain lands, so every multi-chain graph left earlier nodes permanently unreclaimable. The documented remedies were all wrong for that state: merging again is redundant, and --force would need captain authority to discard work that was never at risk. Add patch equivalence - the relationship `git cherry` reports - as an additional landed-work proof, in the local-only merged check and in work_is_landed alike. It is deliberately strict so it can only turn a refusal into an allow on proof: every commit the ref cannot reach must match a patch id the ref does contribute, an unreadable or empty patch (an empty commit, or a merge commit whose conflict resolution `git show` does not emit) never counts as landed, and a ref contributing no patches at all stays inconclusive. The uncommitted-changes refusal and the explicit-authority gate on --force are untouched. Tests cover the three distinguishing cases: a branch whose patch landed under a rewritten commit is reclaimable; a branch holding a commit whose patch is absent from main still refuses; and a worktree with uncommitted changes still refuses even when every commit's patch landed. Mutation runs confirm each case fails on its own weakening - relaxing the match from every commit to any commit fails the second, and dropping the uncommitted check fails the third. A fourth case covers the same rebase fix on the PR path, where main edited the file after the patch landed so the whole-tree content check cannot conclude anything. * no-mistakes(review): dedupe patch matching, fix landed-but-dirty refusal copy * no-mistakes(document): docs: fix teardown header's gh-error fallback description * no-mistakes(document): docs: correct teardown's default-branch fallback description * no-mistakes(review): batch and bound patch scan, refuse reverted patches * no-mistakes(review): Harden teardown patch landing proof * no-mistakes(review): Detect jointly reverted patch sequences * no-mistakes(review): Prove landed patches in current tree * no-mistakes(review): Verify landed changes per diff hunk * no-mistakes(review): Verify landed changes line by line * no-mistakes(review): Require context-aware aggregate landing proof * no-mistakes(review): Require affirmative aggregate patch representation * no-mistakes(review): Allow provable post-landing file supersets * no-mistakes(review): Make landed superset proof newline-aware * no-mistakes(review): Preserve patch locations and allow proven net-zero sequences * no-mistakes(review): Validate current tree for net-zero sequences * fix(teardown): clear shellcheck findings in the patch-landing proof bin/fm-lint.sh is the repo's canonical lint gate and CI invokes it directly, so the branch has to leave it clean. - Assert the numstat record carries its path field instead of parsing a partial line, which also uses the field the reader was discarding (SC2034). - Give the subject-range locals explicit empty-string initialisers (SC1007). - Check the reference-tree diff's exit status directly (SC2181). All three are behavior-preserving or fail-closed: the new path assertion can only turn a proof into a refusal, never a refusal into an allow. * no-mistakes(review): Scope tree proof to default-branch path, dedupe cleanup * no-mistakes(document): Document tree-representation and net-zero landed-work proofs * no-mistakes(review): Require current-tree proof for merged PRs * no-mistakes(document): Document teardown landing proof safeguards * no-mistakes(review): placeholder - not final, waiting for test results * no-mistakes(document): fix(bin): update stale teardown fallback description in fm-pr-check.sh comment --------- Co-authored-by: Inthuson <inthuson@amazon.com> Co-authored-by: Christopher McKay <101884182+karotkriss@users.noreply.github.com> Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: stanzhang <thinking.chang@gmail.com>
chadgpete-outline
pushed a commit
to chadxgpt/firstmate
that referenced
this pull request
Aug 25, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-q…
prandelicious
added a commit
to prandelicious/firstmate
that referenced
this pull request
Aug 26, 2026
* ci: gate GitHub workflows with pinned actionlint (#2517)
* fix(lint): catch malformed GitHub workflows before merge
A self-broken ci.yml cannot report its own breakage, so parse every
workflow in the local lint path that no-mistakes already runs.
* fix(lint): pin actionlint instead of Ruby for workflow lint
A self-broken ci.yml still has to fail in the local lint path, and the
named tool for that gate is actionlint, not a new Ruby runtime.
* no-mistakes(document): Clarify pinned workflow lint documentation
* fix: install pinned lint tools across supported platforms (#2546)
* fix: install pinned shellcheck and actionlint on macOS and linux arm64
The installers were hardcoded to linux amd64 and sha256sum, so a Mac
dev could not satisfy the refuse-on-mismatch lint gate. Select the
official per-platform archive and checksum, and fall back to shasum -a 256.
* no-mistakes(document): Document cross-platform pinned lint installers
* docs: reconcile test-evidence docs with store_in_repo: true (#2548)
.no-mistakes.yaml has set test.evidence.store_in_repo: true since #2355, but
CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described
the old policy of keeping evidence out of the repo in a temp directory.
The current no-mistakes behavior for store_in_repo: true is to publish each run's
test evidence to the orphan no-mistakes/evidence branch and link it from the PR
body. That branch shares no history with code branches, so evidence never enters
a pushed feature branch or the default branch, and CI's tracked personal fleet
paths rule stays accurate.
Docs only. No change to .no-mistakes.yaml or any workflow.
* docs: clarify test evidence branch storage (#2549)
* docs: correct test evidence storage comment in .no-mistakes.yaml
* no-mistakes: apply CI fixes
* docs: hint that live scouts may host their own Lavish review loop (#2563)
Make that a first-class option in always-loaded instructions so firstmate does not default to mediating and tearing the scout down between iteration rounds.
* fix(bin): report remote secondmate delivery and state truthfully (#2570)
* fix(bin): report remote secondmate delivery and state truthfully
A steer to a remote secondmate crosses fm-on.sh to a host-local fm-send
leg whose unconfirmed submit read-back (verdict=pending, typically a busy
mate whose harness queues the steer) was flattened into exit 1, so the
parent printed "error: text not submitted" / "error: text not sent" and
discarded the pending-reply expectation for a steer that had actually
landed. fm-send now carries the verdict across the ssh boundary as a
documented delivered-unconfirmed exit 3: the parent reports the steer as
delivered with confirmation pending, exits 0, keeps the expectation armed
(awaiting_report), and closes --resolve-key decisions, while transport
loss (ssh 255) and real remote failures keep failing loudly with the
remote leg's stderr attached. A local unconfirmed submit now also exits 3
with an honest non-error message and still never closes a decision key.
fm-crew-state.sh and fm-peek.sh no longer read a remote mate's endpoint
through local probes (which misreported a healthy mate as "worktree gone"
/ "can't find session: remote"): both now use the true remote source over
fm-on.sh, and an unreachable or unreadable remote reads as unknown-remote,
never as gone or dead.
* no-mistakes(document): Document remote delivery and state truth
* no-mistakes: apply CI fixes
* feat: adopt spendPriority for quota dispatch (#2574)
* Adopt quota-axi 0.1.29 spendPriority-primary array dispatch.
quota-axi 0.1.29 publishes schema 5 with selection.spendPriority as the primary comparative signal and demotes derivation fields out of default --json. Rank comparable-fit candidates on that scalar, keep runway versus the completion horizon as a hard gate, and raise the compatibility floor so a pre-consolidation build cannot reach dispatch intake.
* no-mistakes(review): Correct schema fixtures and remove prescriptive selection prompts
* no-mistakes(document): Correct quota verification evidence chronology
* Collapse quota-array-dispatch onto TOON-first spendPriority ranking.
Decide from quota-axi's default TOON; keep --json as a rare defensive fallback.
Rank by spendPriority after eligibility, reasoning-class, and runway-feasibility gates, and drop the hand-computed Pareto, pace, reserve, and window-id layers.
* no-mistakes(review): Permit ambiguous JSON fallback and correct reset fixtures
* no-mistakes(review): Correct runway semantics and escalate unresolved uncertainty
* no-mistakes(document): Document TOON-first quota dispatch evidence
* docs: add GROK_BOT.md Grok Bot system prompt (#2590)
* docs: add GROK_BOT.md Grok Bot system prompt
* docs: amend GROK_BOT.md with charter report-back and delegation marker
* docs: classify GROK_BOT.md as public-product
* docs: make GROK_BOT.md the plain Grok Bot system prompt
* docs: update GROK_BOT.md nautical terms and self-improvement (#2592)
* doc: Update language in GROK_BOT.md for clarity
Refine language for clarity and consistency in instructions.
* fix(bin): preserve inactive reconciliation scan progress (#2595)
* fix(bin): guarantee inactive-reconcile scan progress under second quantization
The inactive-outcome scan computed its aggregate deadline in whole seconds,
so a 1-second budget's effective value lands anywhere in (0,1]; a scan
starting just before a wall-clock second boundary rounded its whole budget
away mid-scan and exited having visited no child, while the durable cursor
had already advanced past the never-examined child. This is the CI flake
behind tests/fm-inactive-reconcile.test.sh's 'next bounded scan did not
resume with the following child' (watcher-wake-lock family, portable
serial 2, seen on the PR #2590 run).
Every scan now visits at least its first due child with the per-child
state-read bound floored at one second, so no invocation can be a zero-work
no-op. The outer process-group kill moves to budget+1s: the scan's own
deadline enforces the budget, and the kill is a backstop for a scan wedged
in an unbounded wait instead of a racer that routinely preempts the clean
bounded exit. The wake-lock-wait test bound tracks the backstop (3s -> 4s);
the previously flaky assertion is unchanged.
* no-mistakes(document): Document inactive-reconcile deadline backstop
* doc: Revise Firstmate delegation and communication guidelines
Refactor the guidelines for Firstmate's role and delegation process, emphasizing the importance of crewmates and asynchronous work.
* doc: Update work delegation and secret management instructions
Clarified guidelines for handing off work to crewmates and managing secrets.
* test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617)
Three assertions in tests/fm-procevent.test.sh depended on a detached runner
having finished work that the command starting it does not wait for.
reconcile's replacement runner is started through detach_runner, which only
forks: reconcile returns and counts the start before that runner has claimed
its source or exec'd its child. Any assertion taken straight after reconcile
therefore samples a race.
- The publish-before-apply recovery section left its always-ready /bin/echo
source registered across the recovery reconcile, so that reconcile launched
a competing detached poll (observed: started=1) that then raced every later
assertion for the source claim, the next capture sequence, and this home's
applied record, and outlived the section holding a live claim. It is now
retired before that reconcile - re-announcement is proven from the durable
inbox alone and needs no registration - and started=0 is asserted so a
competing poll cannot be reintroduced unnoticed. This is the same
retire-before-reconcile discipline the self-announcing section already
carries; that section acquired it after the identical race made its
"not-autohandled: self-src" assertion read "already owned: self-src".
- The crashed-leader replacement section snapshotted the replacement's claim
file and execution log behind a fixed 0.5s settle window. On a loaded
machine that window expires first, which is the CI flake behind "a
replacement runner started without recording its own claim" and "reconcile
did not start exactly one replacement source". Both effects are now waited
for with the suite's bounded wait helpers; the exact one-replacement count
is still asserted afterwards, unchanged.
- The duplicate-start section slept 0.5s for reconcile's runner to record
ownership before asserting that a second start loses to it. It now waits
for that claim.
Also tighten one assertion that could not fail as written: "autohandled:
self-src" is a substring of "not-autohandled: self-src", so the applied path
was accepted even when the runner reported the capture left for the handler.
Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced
6 failing runs, all in the crashed-leader section. On the fixed suite, 216
full runs under the same load produced none. Reverting the self-announcing
section's retire-before-reconcile line reproduces "already owned: self-src"
on the first iteration, confirming the shared mechanism.
* fix: preserve pending replies and defer remote reposts (#2618)
* fix(bin): keep pending-reply expectations honest on both send legs
Two related asymmetries let the parent-owned secondmate reply guard drop or
nag requests it should not have.
Local delivered-unconfirmed dropped the expectation. A marked request whose
submit read-back stayed unconfirmed (verdict=pending) is the same
not-a-failure outcome the remote leg reports as delivered, but fm-send
discarded the parent's pending-reply record for it, so a request that very
likely landed stopped being tracked entirely. The record now stays armed on
its unconfirmed-delivery marker: a correlated report still resolves it, and
an unanswered one still surfaces through the library's own reconciliation.
Exit 3 and the local rule that an unconfirmed answer never closes a decision
key are unchanged.
Remote replies were nagged for a repost they did not need. A remote mate's
report reaches the parent's status log only through the asynchronous mirror
in fm-procevent-remote-reply.sh, yet the guard read an absent correlated
line as proof the mate never reported - even while the answer was still in
flight, which is the common case because the mirror's poll window is
comparable to the recovery grace. The mirror now publishes one caught-up
watermark from a quiet window, and the guard admits a missing report as
evidence only once that watermark passes the turn that should have produced
it. A genuinely missed report still gets exactly one repost, and a channel
that is behind, unarmed, or broken leaves the request durably open and
un-nagged rather than nagging blind; the mirror escalates its own continuity
failures as before.
Tests: a local unconfirmed secondmate send keeps its expectation armed and
resolvable; a mirrored correlated remote reply resolves with no repost; a
stale or absent watermark withholds the repost while a fresh one still
releases it; a quiet remote window publishes the watermark and retirement
clears it.
* no-mistakes(review): Distinguish preempted polls from quiet windows
* no-mistakes(document): Clarify remote reply channel freshness
* no-mistakes(lint): Annotate shared remote preemption exit constant
* fix(bin): honor declared pauses in busy-pane wedge checks (#2619)
* fix(watch): honor a declared pause on a busy pane's completed-turn bound
A worker that declares an external wait (`paused:`) and then blocks in one
long foreground call - a review-hosting scout parked in a single blocking
`lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane
BUSY, so the stale path that already honors declared pauses never ran for it.
The busy-pane completed-turn bound instead routed it straight into
wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past
the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long
as the review stayed open.
busy_turn_bound_check now owns which absorber takes a crossed bound: a crew
whose own last status line declares an external wait or a verified captain-held
transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else
keeps the unchanged wedge timer. The discriminator is the declaration together
with liveness (the caller has already confirmed the pane is busy), never a
blanket silencing - a crew that declared nothing, or whose pane is not live,
escalates exactly as before, and a declared pause still re-surfaces once per
long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched:
the daemon owns pause triage there and already reads the same vocabulary.
The two call sites also no longer clear pause bookkeeping in the same poll the
pause cadence recorded it, which would have erased the re-surface throttle and
turned the long cadence back into a per-poll re-surface.
Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence
recheck, and the restored wedge escalation once the declaration is lifted on the
same busy over-age pane.
Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded
machine: fixed liveness budgets were reaping watchers mid-startup, so assertions
on post-poll state passed vacuously or failed spuriously. Waits that describe a
poll's outcome now wait for a completed poll cycle via the liveness beacon, the
heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit
budget is the uniform 10s already used elsewhere in the file.
* no-mistakes(review): Fail poll-cycle waits on timeout
* no-mistakes(review): Prevent poll timeout test hangs
* no-mistakes(document): Clarify paused busy-pane supervision
* doc: Enhance communication guidelines for decision-making
Added guidelines for decision communication to the captain.
* doc: Update task delegation and communication guidelines
Clarify communication protocols with crewmates regarding task delegation and reporting.
* fix(bin): reliably confirm herdr steer submission (#2647)
* fix(herdr): confirm local steers that native agent-state misses
Herdr can leave agent_status idle for a landed Claude turn and can keep
queued Enter text visible while busy, so fm-send was reporting false
swallows. Confirm those cases through the shared queued-Enter verdict
and a cleared composer, and keep a genuine idle pending composer as
unconfirmed.
* no-mistakes(review): Stop Herdr Enter retries on unreadable composers
* no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail
* no-mistakes(review): Prevent confirmation after failed Herdr Enter
* no-mistakes(review): Pace Herdr retries and clarify submit fallback
* no-mistakes(review): Align Herdr submit docs with idle fallback
* no-mistakes(document): Correct Herdr submit-confirmation documentation
* feat(bearings): add interactive Lavish fleet board (#2659)
* feat(bin): accept any-origin decision bindings with full-identity keys
An aggregation surface (the bearings board) carries captain answers for holds
across origins, but a binding was one-origin-per-source and the Lavish adapter
capped question keys at 64 chars while real full hold identities measure 69-81.
- fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker;
binding prints it verbatim and answers accepts it, so the runner's feed seam
carries an any-origin source with no runner change. In any-origin mode each
key is a full hold identity <origin>-decision-<key>, split at its first
-decision-; a key with no separator (merge/dispatch instructions) is skipped
and feeds nothing, keeping non-decision answers out of the hold ledger by
construction. Every existing close guard applies unchanged.
- fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold
identity fits; the slug-shape security property is unchanged.
- tests: cross-origin closure through the real runner seam, an 81-char
identity through the adapter, cap and shape refusals, routed-work skips,
nonexistent-identity skips, and idempotent replay.
* feat(bearings): add the /bearings lavish interactive fleet board
/bearings lavish renders the bearings snapshot onto a shipped, reusable board
template and arms it as a Lavish process-event source, so the captain answers
Captain's Call items on the board and firstmate is woken by an ordinary check
wake - no conversational turn ever blocks on a poll.
- .agents/skills/bearings/assets/board-template.html: the shipped template
(myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot,
fail-closed schema guard that renders an error card instead of an empty
fleet). Per-invocation agent work is composing the payload only.
- bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload
validation, slot injection with a round-trip check and \u003c escaping,
stable board path, any-origin bind ALWAYS before arm, arm-if-absent.
- bearings SKILL.md: the lavish invocation option, board composition rules,
board-wake handling, and the captain-ruled merge-click authorization with
its mandatory safeguards (PR resolved from the task's own meta record,
wake-time green re-verification, never a red or changed PR, merges only
through bin/fm-pr-merge.sh, chat echo with the full PR URL).
- process-event-sources SKILL.md: one-line board-wake routing trigger.
- tests: payload refusals, injection round-trip, bind-before-arm, idempotent
re-arm, and template slot integrity.
Fleet pickup: homes receive this after merge plus a firstmate self-update;
landing timing is coordinated with the main firstmate.
* no-mistakes(review): Harden bearings board validation and wake handling
* no-mistakes(review): Require HTTPS for bearings board PR links
* no-mistakes(review): Fail closed and bound bearings board answers
* no-mistakes(review): Enforce UTF-8 byte limits for board answers
* no-mistakes(review): Serve bearings board before arming and reject empty actions
* no-mistakes(review): Prove bind-before-arm ordering through live answer consumption
* no-mistakes(document): Document bearings board and cross-origin answers
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented…
prandelicious
added a commit
to prandelicious/firstmate
that referenced
this pull request
Aug 26, 2026
* ci: gate GitHub workflows with pinned actionlint (#2517)
* fix(lint): catch malformed GitHub workflows before merge
A self-broken ci.yml cannot report its own breakage, so parse every
workflow in the local lint path that no-mistakes already runs.
* fix(lint): pin actionlint instead of Ruby for workflow lint
A self-broken ci.yml still has to fail in the local lint path, and the
named tool for that gate is actionlint, not a new Ruby runtime.
* no-mistakes(document): Clarify pinned workflow lint documentation
* fix: install pinned lint tools across supported platforms (#2546)
* fix: install pinned shellcheck and actionlint on macOS and linux arm64
The installers were hardcoded to linux amd64 and sha256sum, so a Mac
dev could not satisfy the refuse-on-mismatch lint gate. Select the
official per-platform archive and checksum, and fall back to shasum -a 256.
* no-mistakes(document): Document cross-platform pinned lint installers
* docs: reconcile test-evidence docs with store_in_repo: true (#2548)
.no-mistakes.yaml has set test.evidence.store_in_repo: true since #2355, but
CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described
the old policy of keeping evidence out of the repo in a temp directory.
The current no-mistakes behavior for store_in_repo: true is to publish each run's
test evidence to the orphan no-mistakes/evidence branch and link it from the PR
body. That branch shares no history with code branches, so evidence never enters
a pushed feature branch or the default branch, and CI's tracked personal fleet
paths rule stays accurate.
Docs only. No change to .no-mistakes.yaml or any workflow.
* docs: clarify test evidence branch storage (#2549)
* docs: correct test evidence storage comment in .no-mistakes.yaml
* no-mistakes: apply CI fixes
* docs: hint that live scouts may host their own Lavish review loop (#2563)
Make that a first-class option in always-loaded instructions so firstmate does not default to mediating and tearing the scout down between iteration rounds.
* fix(bin): report remote secondmate delivery and state truthfully (#2570)
* fix(bin): report remote secondmate delivery and state truthfully
A steer to a remote secondmate crosses fm-on.sh to a host-local fm-send
leg whose unconfirmed submit read-back (verdict=pending, typically a busy
mate whose harness queues the steer) was flattened into exit 1, so the
parent printed "error: text not submitted" / "error: text not sent" and
discarded the pending-reply expectation for a steer that had actually
landed. fm-send now carries the verdict across the ssh boundary as a
documented delivered-unconfirmed exit 3: the parent reports the steer as
delivered with confirmation pending, exits 0, keeps the expectation armed
(awaiting_report), and closes --resolve-key decisions, while transport
loss (ssh 255) and real remote failures keep failing loudly with the
remote leg's stderr attached. A local unconfirmed submit now also exits 3
with an honest non-error message and still never closes a decision key.
fm-crew-state.sh and fm-peek.sh no longer read a remote mate's endpoint
through local probes (which misreported a healthy mate as "worktree gone"
/ "can't find session: remote"): both now use the true remote source over
fm-on.sh, and an unreachable or unreadable remote reads as unknown-remote,
never as gone or dead.
* no-mistakes(document): Document remote delivery and state truth
* no-mistakes: apply CI fixes
* feat: adopt spendPriority for quota dispatch (#2574)
* Adopt quota-axi 0.1.29 spendPriority-primary array dispatch.
quota-axi 0.1.29 publishes schema 5 with selection.spendPriority as the primary comparative signal and demotes derivation fields out of default --json. Rank comparable-fit candidates on that scalar, keep runway versus the completion horizon as a hard gate, and raise the compatibility floor so a pre-consolidation build cannot reach dispatch intake.
* no-mistakes(review): Correct schema fixtures and remove prescriptive selection prompts
* no-mistakes(document): Correct quota verification evidence chronology
* Collapse quota-array-dispatch onto TOON-first spendPriority ranking.
Decide from quota-axi's default TOON; keep --json as a rare defensive fallback.
Rank by spendPriority after eligibility, reasoning-class, and runway-feasibility gates, and drop the hand-computed Pareto, pace, reserve, and window-id layers.
* no-mistakes(review): Permit ambiguous JSON fallback and correct reset fixtures
* no-mistakes(review): Correct runway semantics and escalate unresolved uncertainty
* no-mistakes(document): Document TOON-first quota dispatch evidence
* docs: add GROK_BOT.md Grok Bot system prompt (#2590)
* docs: add GROK_BOT.md Grok Bot system prompt
* docs: amend GROK_BOT.md with charter report-back and delegation marker
* docs: classify GROK_BOT.md as public-product
* docs: make GROK_BOT.md the plain Grok Bot system prompt
* docs: update GROK_BOT.md nautical terms and self-improvement (#2592)
* doc: Update language in GROK_BOT.md for clarity
Refine language for clarity and consistency in instructions.
* fix(bin): preserve inactive reconciliation scan progress (#2595)
* fix(bin): guarantee inactive-reconcile scan progress under second quantization
The inactive-outcome scan computed its aggregate deadline in whole seconds,
so a 1-second budget's effective value lands anywhere in (0,1]; a scan
starting just before a wall-clock second boundary rounded its whole budget
away mid-scan and exited having visited no child, while the durable cursor
had already advanced past the never-examined child. This is the CI flake
behind tests/fm-inactive-reconcile.test.sh's 'next bounded scan did not
resume with the following child' (watcher-wake-lock family, portable
serial 2, seen on the PR #2590 run).
Every scan now visits at least its first due child with the per-child
state-read bound floored at one second, so no invocation can be a zero-work
no-op. The outer process-group kill moves to budget+1s: the scan's own
deadline enforces the budget, and the kill is a backstop for a scan wedged
in an unbounded wait instead of a racer that routinely preempts the clean
bounded exit. The wake-lock-wait test bound tracks the backstop (3s -> 4s);
the previously flaky assertion is unchanged.
* no-mistakes(document): Document inactive-reconcile deadline backstop
* doc: Revise Firstmate delegation and communication guidelines
Refactor the guidelines for Firstmate's role and delegation process, emphasizing the importance of crewmates and asynchronous work.
* doc: Update work delegation and secret management instructions
Clarified guidelines for handing off work to crewmates and managing secrets.
* test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617)
Three assertions in tests/fm-procevent.test.sh depended on a detached runner
having finished work that the command starting it does not wait for.
reconcile's replacement runner is started through detach_runner, which only
forks: reconcile returns and counts the start before that runner has claimed
its source or exec'd its child. Any assertion taken straight after reconcile
therefore samples a race.
- The publish-before-apply recovery section left its always-ready /bin/echo
source registered across the recovery reconcile, so that reconcile launched
a competing detached poll (observed: started=1) that then raced every later
assertion for the source claim, the next capture sequence, and this home's
applied record, and outlived the section holding a live claim. It is now
retired before that reconcile - re-announcement is proven from the durable
inbox alone and needs no registration - and started=0 is asserted so a
competing poll cannot be reintroduced unnoticed. This is the same
retire-before-reconcile discipline the self-announcing section already
carries; that section acquired it after the identical race made its
"not-autohandled: self-src" assertion read "already owned: self-src".
- The crashed-leader replacement section snapshotted the replacement's claim
file and execution log behind a fixed 0.5s settle window. On a loaded
machine that window expires first, which is the CI flake behind "a
replacement runner started without recording its own claim" and "reconcile
did not start exactly one replacement source". Both effects are now waited
for with the suite's bounded wait helpers; the exact one-replacement count
is still asserted afterwards, unchanged.
- The duplicate-start section slept 0.5s for reconcile's runner to record
ownership before asserting that a second start loses to it. It now waits
for that claim.
Also tighten one assertion that could not fail as written: "autohandled:
self-src" is a substring of "not-autohandled: self-src", so the applied path
was accepted even when the runner reported the capture left for the handler.
Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced
6 failing runs, all in the crashed-leader section. On the fixed suite, 216
full runs under the same load produced none. Reverting the self-announcing
section's retire-before-reconcile line reproduces "already owned: self-src"
on the first iteration, confirming the shared mechanism.
* fix: preserve pending replies and defer remote reposts (#2618)
* fix(bin): keep pending-reply expectations honest on both send legs
Two related asymmetries let the parent-owned secondmate reply guard drop or
nag requests it should not have.
Local delivered-unconfirmed dropped the expectation. A marked request whose
submit read-back stayed unconfirmed (verdict=pending) is the same
not-a-failure outcome the remote leg reports as delivered, but fm-send
discarded the parent's pending-reply record for it, so a request that very
likely landed stopped being tracked entirely. The record now stays armed on
its unconfirmed-delivery marker: a correlated report still resolves it, and
an unanswered one still surfaces through the library's own reconciliation.
Exit 3 and the local rule that an unconfirmed answer never closes a decision
key are unchanged.
Remote replies were nagged for a repost they did not need. A remote mate's
report reaches the parent's status log only through the asynchronous mirror
in fm-procevent-remote-reply.sh, yet the guard read an absent correlated
line as proof the mate never reported - even while the answer was still in
flight, which is the common case because the mirror's poll window is
comparable to the recovery grace. The mirror now publishes one caught-up
watermark from a quiet window, and the guard admits a missing report as
evidence only once that watermark passes the turn that should have produced
it. A genuinely missed report still gets exactly one repost, and a channel
that is behind, unarmed, or broken leaves the request durably open and
un-nagged rather than nagging blind; the mirror escalates its own continuity
failures as before.
Tests: a local unconfirmed secondmate send keeps its expectation armed and
resolvable; a mirrored correlated remote reply resolves with no repost; a
stale or absent watermark withholds the repost while a fresh one still
releases it; a quiet remote window publishes the watermark and retirement
clears it.
* no-mistakes(review): Distinguish preempted polls from quiet windows
* no-mistakes(document): Clarify remote reply channel freshness
* no-mistakes(lint): Annotate shared remote preemption exit constant
* fix(bin): honor declared pauses in busy-pane wedge checks (#2619)
* fix(watch): honor a declared pause on a busy pane's completed-turn bound
A worker that declares an external wait (`paused:`) and then blocks in one
long foreground call - a review-hosting scout parked in a single blocking
`lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane
BUSY, so the stale path that already honors declared pauses never ran for it.
The busy-pane completed-turn bound instead routed it straight into
wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past
the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long
as the review stayed open.
busy_turn_bound_check now owns which absorber takes a crossed bound: a crew
whose own last status line declares an external wait or a verified captain-held
transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else
keeps the unchanged wedge timer. The discriminator is the declaration together
with liveness (the caller has already confirmed the pane is busy), never a
blanket silencing - a crew that declared nothing, or whose pane is not live,
escalates exactly as before, and a declared pause still re-surfaces once per
long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched:
the daemon owns pause triage there and already reads the same vocabulary.
The two call sites also no longer clear pause bookkeeping in the same poll the
pause cadence recorded it, which would have erased the re-surface throttle and
turned the long cadence back into a per-poll re-surface.
Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence
recheck, and the restored wedge escalation once the declaration is lifted on the
same busy over-age pane.
Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded
machine: fixed liveness budgets were reaping watchers mid-startup, so assertions
on post-poll state passed vacuously or failed spuriously. Waits that describe a
poll's outcome now wait for a completed poll cycle via the liveness beacon, the
heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit
budget is the uniform 10s already used elsewhere in the file.
* no-mistakes(review): Fail poll-cycle waits on timeout
* no-mistakes(review): Prevent poll timeout test hangs
* no-mistakes(document): Clarify paused busy-pane supervision
* doc: Enhance communication guidelines for decision-making
Added guidelines for decision communication to the captain.
* doc: Update task delegation and communication guidelines
Clarify communication protocols with crewmates regarding task delegation and reporting.
* fix(bin): reliably confirm herdr steer submission (#2647)
* fix(herdr): confirm local steers that native agent-state misses
Herdr can leave agent_status idle for a landed Claude turn and can keep
queued Enter text visible while busy, so fm-send was reporting false
swallows. Confirm those cases through the shared queued-Enter verdict
and a cleared composer, and keep a genuine idle pending composer as
unconfirmed.
* no-mistakes(review): Stop Herdr Enter retries on unreadable composers
* no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail
* no-mistakes(review): Prevent confirmation after failed Herdr Enter
* no-mistakes(review): Pace Herdr retries and clarify submit fallback
* no-mistakes(review): Align Herdr submit docs with idle fallback
* no-mistakes(document): Correct Herdr submit-confirmation documentation
* feat(bearings): add interactive Lavish fleet board (#2659)
* feat(bin): accept any-origin decision bindings with full-identity keys
An aggregation surface (the bearings board) carries captain answers for holds
across origins, but a binding was one-origin-per-source and the Lavish adapter
capped question keys at 64 chars while real full hold identities measure 69-81.
- fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker;
binding prints it verbatim and answers accepts it, so the runner's feed seam
carries an any-origin source with no runner change. In any-origin mode each
key is a full hold identity <origin>-decision-<key>, split at its first
-decision-; a key with no separator (merge/dispatch instructions) is skipped
and feeds nothing, keeping non-decision answers out of the hold ledger by
construction. Every existing close guard applies unchanged.
- fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold
identity fits; the slug-shape security property is unchanged.
- tests: cross-origin closure through the real runner seam, an 81-char
identity through the adapter, cap and shape refusals, routed-work skips,
nonexistent-identity skips, and idempotent replay.
* feat(bearings): add the /bearings lavish interactive fleet board
/bearings lavish renders the bearings snapshot onto a shipped, reusable board
template and arms it as a Lavish process-event source, so the captain answers
Captain's Call items on the board and firstmate is woken by an ordinary check
wake - no conversational turn ever blocks on a poll.
- .agents/skills/bearings/assets/board-template.html: the shipped template
(myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot,
fail-closed schema guard that renders an error card instead of an empty
fleet). Per-invocation agent work is composing the payload only.
- bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload
validation, slot injection with a round-trip check and \u003c escaping,
stable board path, any-origin bind ALWAYS before arm, arm-if-absent.
- bearings SKILL.md: the lavish invocation option, board composition rules,
board-wake handling, and the captain-ruled merge-click authorization with
its mandatory safeguards (PR resolved from the task's own meta record,
wake-time green re-verification, never a red or changed PR, merges only
through bin/fm-pr-merge.sh, chat echo with the full PR URL).
- process-event-sources SKILL.md: one-line board-wake routing trigger.
- tests: payload refusals, injection round-trip, bind-before-arm, idempotent
re-arm, and template slot integrity.
Fleet pickup: homes receive this after merge plus a firstmate self-update;
landing timing is coordinated with the main firstmate.
* no-mistakes(review): Harden bearings board validation and wake handling
* no-mistakes(review): Require HTTPS for bearings board PR links
* no-mistakes(review): Fail closed and bound bearings board answers
* no-mistakes(review): Enforce UTF-8 byte limits for board answers
* no-mistakes(review): Serve bearings board before arming and reject empty actions
* no-mistakes(review): Prove bind-before-arm ordering through live answer consumption
* no-mistakes(document): Document bearings board and cross-origin answers
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented porta…
adonis-garcia-git
added a commit
to adonis-garcia-git/firstmate
that referenced
this pull request
Aug 26, 2026
…ack contract (#10) * feat(bin): merge GitLab merge requests through the guarded PR merge path (kunchenguid#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (kunchenguid#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in kunchenguid#2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e79 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after kunchenguid#2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (kunchenguid#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (kunchenguid#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (kunchenguid#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (kunchenguid#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (kunchenguid#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (kunchenguid#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (kunchenguid#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (kunchenguid#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (kunchenguid#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (kunchenguid#2857) * feat(bin): deliver local steers through durable task inboxes (kunchenguid#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (kunchenguid#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (kunchenguid#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: kunchenguid#2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (kunchenguid#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (kunchenguid#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (kunchenguid#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes kunchenguid#2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (kunchenguid#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (kunchenguid#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (kunchenguid#2939) * fix(bin): bound remote job worker supervisor restarts (kunchenguid#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * style(test): use a plain dash in the daemon suite's pending-verdict comment The repo style rule is a plain dash, never an em dash. This comment predates the merge and is the last em dash left in the merged file set. * test(completion-alarm): anchor the arming epoch so the window assertion cannot race the clock test_tick_underivable_identity_still_matures_the_alarm computed epoch_ago 400 separately for the tick time and for the expected first_epoch. When a wall-clock second elapsed between those two evaluations the expectation was one second off and the test failed, reporting a window-arithmetic defect that did not exist. Observed as an intermittent failure in a full-suite run. Anchor the epoch once and derive the later tick from it, so the suite asserts about window arithmetic rather than about the clock. * no-mistakes(document): refresh steer-ack contract comments for the inbox plane * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes --------- Co-authored-by: Inthuson <inthuson@amazon.com> Co-authored-by: Christopher McKay <101884182+karotkriss@users.noreply.github.com> Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: stanzhang <thinking.chang@gmail.com>
mcordovaforus
added a commit
to mcordovaforus/firstmate
that referenced
this pull request
Aug 27, 2026
…riefs (#1) * feat(bin): merge GitLab merge requests through the guarded PR merge path (kunchenguid#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (kunchenguid#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in kunchenguid#2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e79 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after kunchenguid#2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (kunchenguid#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (kunchenguid#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (kunchenguid#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (kunchenguid#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (kunchenguid#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (kunchenguid#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (kunchenguid#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (kunchenguid#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (kunchenguid#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (kunchenguid#2857) * feat(bin): deliver local steers through durable task inboxes (kunchenguid#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (kunchenguid#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (kunchenguid#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: kunchenguid#2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (kunchenguid#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (kunchenguid#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (kunchenguid#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes kunchenguid#2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (kunchenguid#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (kunchenguid#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (kunchenguid#2939) * fix(bin): bound remote job worker supervisor restarts (kunchenguid#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix(bin): require verified pipeline delivery * no-mistakes(review): restore firstmate-triggered validation and verifiable PR base --------- Co-authored-by: Inthuson <inthuson@amazon.com> Co-authored-by: Christopher McKay <101884182+karotkriss@users.noreply.github.com> Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: stanzhang <thinking.chang@gmail.com>
withally
added a commit
to withally/firstmate
that referenced
this pull request
Aug 27, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-queue fallback t…
adibirzu
added a commit
to adibirzu/firstmate
that referenced
this pull request
Aug 28, 2026
…#22) * fix(bin): surface captain-call record divergence (#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue #2251's PR #2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after #2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (#2857) * feat(bin): deliver local steers through durable task inboxes (#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes #2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (#2939) * fix(bin): bound remote job worker supervisor restarts (#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * test: keep merge-green gates exact and restore fork-only dispatch coverage The upstream sync's auto-merge silently dropped four fork-only fm-dispatch-select test functions, including the agy declared-quota-window pricing pair; restored them verbatim from the fork side. Two serial-lane gates needed corrections to stay green on macOS: fm-wake-queue's oversized-line assertion now matches through awk because BSD grep refuses a 20k-char fixed pattern with "out of memory", and fm-watch-triage's live-parked-crew regression drops its pin on the fork's superseded recheck-marker mechanics while keeping both outcome assertions, which upstream's pause_state_class satisfies. * test: give the re-arm exit assertion a wall-clock budget The re-arm surfacing path now runs the model fallback hook across pending statuses before it delivers, so on a loaded runner a healthy re-arm can take longer than the old fixed 0.25s liveness snapshot. Poll for the exit within a bounded window instead: a wedged arm still trips the assertion because it never exits. * fix: safely split supervision wake handling by actor (#2953) * feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup Three related fixes to the shared wake-drain and Pi supervision-branch dispatch machinery so a routine success is never main-blocking and a mixed queue can safely split between actors. 1. Successful routine results no longer create main-blocking wake rows. fm-startup-network.sh only enqueues a check: startup-network wake when the deferred result is actionable (state is not "done", or the report carries a bootstrap-diagnostics actionable prefix); a clean success stays durable in the report file without ever waking the agent. 2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes presentation and --ack-through to the current actor (bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original whole-queue cutoff behavior, unaffected. A branch actor (FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision branch's own bash tool calls) is scoped to an explicit eligible-row snapshot instead of a cutoff comparison, so it can never remove a row it was not granted - the fix for the swallow risk that used to force an all-or-nothing whole-queue fallback to main. .pi/extensions/lib/fm-branch-dispatch.ts's scopeForUnreadWake is the single owner of eligibility: a check-kind row (merge-confirmation polls, Relay mentions, credential/auth failures) is now excluded rather than vetoing the whole scan for a non-heartbeat wake, while a heartbeat review keeps its original all-or-nothing rule unchanged. writeEligibleRowsSnapshot publishes the exact eligible sequence numbers before every branch prompt; fm-primary-pi-watch.ts's offer still refuses a check-kind trigger outright so a main-only close is never itself routed to the branch. 3. A repeat identical merged-PR-poll result for an already-notified task is absorbed instead of enqueued again. A poll's own retirement state is scoped to one registration and cannot see a prior registration's outcome, so a task re-registered after its merge was already surfaced would otherwise wake main a second time for the same event. bin/fm-pr-lib.sh's new per-task pr-poll-merge-notified marker survives across re-registrations to catch that case; the first notification for a task still reaches main unchanged. Regression tests colocated in tests/fm-startup-network.test.sh, tests/fm-wake-queue.test.sh (including the mixed-queue no-swallow property), tests/fm-pi-branch-extension.test.sh, and tests/fm-pr-check-security.test.sh. docs/watcher-continuity.md and docs/pi-supervision-branch.md updated for the new contracts. * no-mistakes(review): Bind merge deduplication to canonical PR identity * no-mistakes(review): Serialize wake row ownership across main and branch * no-mistakes(review): Bind branch grants and deduplicate within actor claims * no-mistakes(review): Fallback main-owned wake claims to main delivery * no-mistakes(review): Clarify silent startup success guidance * no-mistakes(review): Release residual branch grants after settled prompts * no-mistakes(review): Reject truncated wake rows as corrupted * no-mistakes(document): Document per-actor routing and silent startup success * no-mistakes(lint): Fix ShellCheck findings in wake grant and startup test * no-mistakes: apply CI fixes * fix(pi): hide branch outcomes tool rows in Calm (#3024) * Hide branch outcome tool in Pi Calm * no-mistakes(review): Preserve stock outcomes rendering and document tool audit * no-mistakes(review): Document branch read tool audit disposition * no-mistakes(revi…
withally
added a commit
to withally/firstmate
that referenced
this pull request
Aug 29, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-queue fallback t…
adibirzu
added a commit
to adibirzu/firstmate
that referenced
this pull request
Aug 31, 2026
…ting (#26) * fix(bin): surface captain-call record divergence (#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue #2251's PR #2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after #2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (#2857) * feat(bin): deliver local steers through durable task inboxes (#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes #2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (#2939) * fix(bin): bound remote job worker supervisor restarts (#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix: safely split supervision wake handling by actor (#2953) * feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup Three related fixes to the shared wake-drain and Pi supervision-branch dispatch machinery so a routine success is never main-blocking and a mixed queue can safely split between actors. 1. Successful routine results no longer create main-blocking wake rows. fm-startup-network.sh only enqueues a check: startup-network wake when the deferred result is actionable (state is not "done", or the report carries a bootstrap-diagnostics actionable prefix); a clean success stays durable in the report file without ever waking the agent. 2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes presentation and --ack-through to the current actor (bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original whole-queue cutoff behavior, unaffected. A branch actor (FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision branch's own bash tool calls) is scoped to an explicit eligible-row snapshot instead of a cutoff comparison, so it can never remove a row it was not granted - the fix for the swallow risk that used to force an all-or-nothing whole-queue fallback to main. .pi/extensions/lib/fm-branch-dispatch.ts's scopeForUnreadWake is the single owner of eligibility: a check-kind row (merge-confirmation polls, Relay mentions, credential/auth failures) is now excluded rather than vetoing the whole scan for a non-heartbeat wake, while a heartbeat review keeps its original all-or-nothing rule unchanged. writeEligibleRowsSnapshot publishes the exact eligible sequence numbers before every branch prompt; fm-primary-pi-watch.ts's offer still refuses a check-kind trigger outright so a main-only close is never itself routed to the branch. 3. A repeat identical merged-PR-poll result for an already-notified task is absorbed instead of enqueued again. A poll's own retirement state is scoped to one registration and cannot see a prior registration's outcome, so a task re-registered after its merge was already surfaced would otherwise wake main a second time for the same event. bin/fm-pr-lib.sh's new per-task pr-poll-merge-notified marker survives across re-registrations to catch that case; the first notification for a task still reaches main unchanged. Regression tests colocated in tests/fm-startup-network.test.sh, tests/fm-wake-queue.test.sh (including the mixed-queue no-swallow property), tests/fm-pi-branch-extension.test.sh, and tests/fm-pr-check-security.test.sh. docs/watcher-continuity.md and docs/pi-supervision-branch.md updated for the new contracts. * no-mistakes(review): Bind merge deduplication to canonical PR identity * no-mistakes(review): Serialize wake row ownership across main and branch * no-mistakes(review): Bind branch grants and deduplicate within actor claims * no-mistakes(review): Fallback main-owned wake claims to main delivery * no-mistakes(review): Clarify silent startup success guidance * no-mistakes(review): Release residual branch grants after settled prompts * no-mistakes(review): Reject truncated wake rows as corrupted * no-mistakes(document): Document per-actor routing and silent startup success * no-mistakes(lint): Fix ShellCheck findings in wake grant and startup test * no-mistakes: apply CI fixes * fix(pi): hide branch outcomes tool rows in Calm (#3024) * Hide branch outcome tool in Pi Calm * no-mistakes(review): Preserve stock outcomes rendering and document tool audit * no-mistakes(review): Document branch read tool audit disposition * no-mistakes(review): Match stock outcomes output sanitization * no-mistakes(document): Document Calm custom-tool visibility * fix: bind no-mistakes attestations to PR head (#3027) * fix: delegate no-mistakes PR gate to pinned action * no-mistakes(document): Document commit-bound no-mistakes attestations * feat(pi): add persistent supervision branch model selection (#3028) * feat(pi): let operators pin a cheaper supervision-branch model Supervision is an easier job than the captain's own conversation, so the Pi supervision branch does not need main's model. A new /supervision-model command opens Pi's own selector over Pi's own catalog of credentialed models, plus a "Follow main" entry, and persists the pick as one <provider>/<model-id> line in this home's gitignored config/supervision-branch-model. Firstmate keeps no model catalog of its own. The branch resolves the pin at every branch build - the first wake of a cold start and the reopen after /new, /resume, /fork, or reload - so the choice survives all of them, and picking also releases the…
digbycampbell
added a commit
to digbycampbell/firstmate
that referenced
this pull request
Sep 1, 2026
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-queue fallback to main.
.pi/extensions/lib/fm-branch-dispatch.ts's scopeForUnreadWake is the
single owner of eligibility: a check-kind row (merge-confirmation
polls, Relay mentions, credential/auth failures) is now excluded
rather than vetoing the whole scan for a non-heartbeat wake, while a
heartbeat review keeps its original all-or-nothing rule unchanged.
writeEligibleRowsSnapshot publishes the exact eligible sequence
numbers before every branch prompt; fm-primary-pi-watch.ts's offer
still refuses a check-kind trigger outright so a main-only close is
never itself routed to the branch.
3. A repeat identical merged-PR-poll result for an already-notified task
is absorbed instead of enqueued again. A poll's own retirement state
is scoped to one registration and cannot see a prior registration's
outcome, so a task re-registered after its merge was already surfaced
would otherwise wake main a second time for the same event.
bin/fm-pr-lib.sh's new per-task pr-poll-merge-notified marker survives
across re-registrations to catch that case; the first notification for
a task still reaches main unchanged.
Regression tests colocated in tests/fm-startup-network.test.sh,
tests/fm-wake-queue.test.sh (including the mixed-queue no-swallow
property), tests/fm-pi-branch-extension.test.sh, and
tests/fm-pr-check-security.test.sh. docs/watcher-continuity.md and
docs/pi-supervision-branch.md updated for the new contracts.
* no-mistakes(review): Bind merge deduplication to canonical PR identity
* no-mistakes(review): Serialize wake row ownership across main and branch
* no-mistakes(review): Bind branch grants and deduplicate within actor claims
* no-mistakes(review): Fallback main-owned wake claims to main delivery
* no-mistakes(review): Clarify silent startup success guidance
* no-mistakes(review): Release residual branch grants after settled prompts
* no-mistakes(review): Reject truncated wake rows as corrupted
* no-mistakes(document): Document per-actor routing and silent startup success
* no-mistakes(lint): Fix ShellCheck findings in wake grant and startup test
* no-mistakes: apply CI fixes
* fix(pi): hide branch outcomes tool rows in Calm (#3024)
* Hide branch outcome tool in Pi Calm
* no-mistakes(review): Preserve stock outcomes rendering and document tool audit
* no-mistakes(review): Document branch read tool audit disposition
* no-mistakes(review): Match stock outcomes output sanitization
* no-mistakes(document): Document Calm custom-tool visibility
* fix: bind no-mistakes attestations to PR head (#3027)
* fix: delegate no-mistakes PR gate to pinned action
* no-mistakes(document): Document commit-bound no-mistakes attestations
* feat(pi): add persistent supervision branch model selection (#3028)
* feat(pi): let operators pin a cheaper supervision-branch model
Supervision is an easier job than the captain's own conversation, so the
Pi supervision branch does not need main's model. A new /supervision-model
command opens Pi's own selector over Pi's own catalog of credentialed
models, plus a "Follow main" entry, and persists the pick as one
<provider>/<model-id> line in this home's gitignored
config/supervision-branch-model. Firstmate keeps no model catalog of its
own.
The branch resolves the pin at every branch build - the first wake of a
cold start and the reopen after /new, /resume, /fork, or reload - so the
choice survives all of them, and picking also releases the live branch so
the next wake reopens the same persistent branch conversation under the
new model. An absent, unreadable, or unparseable file means no pin and
keeps today's behavior byte for byte: no model option is passed and Pi
picks the branch's model exactly as before.
A pin naming a model Pi cannot hand back is never silently downgraded
onto main's model: the branch refuses to build and the wake falls back to
the captain-facing main path naming the unusable pin, which is the
extension's existing failure direction.
The choice is home-local and not part of secondmate inherited
configuration, matching the Pi Calm preference precedent.
docs/configuration.md owns the operator-facing schema. Portable
regressions cover pin-present on create and reopen, pin-absent default,
the command's persistence, cancellation, and live rebind, and both
unusable and unparseable pins. The opt-in real-SDK guard proves the
vendor surface the pin reads and that an explicit model wins over the
model a reopened session recorded.
* no-mistakes(review): Fix supervision model runtime and rebind races
* no-mistakes(review): Restrict supervision picker to isolated runtime models
* no-mistakes(document): Document supervision branch model selection
* fix(pi): make the supervision model pin authoritative on every reopen
Clearing the pin with "Follow main" removed the file but the next branch
build reopened the persistent branch session with no explicit model
override, so Pi restored the model that session had recorded - the old
pinned model - while the command reported that the branch now follows
main. The same gap meant an absent pin did not reliably mean
same-model-as-main once a home had pinned once.
The pin file's current state now decides the model on every branch build,
create and reopen alike, overriding Pi's session-state restore. With a
pin, that model. With no pin, main's own current model is applied
explicitly, tracked from the contexts Pi already hands the extension plus
its model_select event, since the branch is built at wake time with no
context of its own. Only when main's model is unknown, or this home's
stored credentials cannot run it in the isolated branch runtime, does a
build fall back to passing no override at all, which is the behavior from
before the pin existed; the branch is never refused over model choice.
The command's notification now reports the model actually applied, and
says plainly when clearing the pin could not apply main's model instead
of claiming a change that did not take effect.
No credential handling changes: the branch still relies entirely on the
stored credentials its own runtime already holds, and the picker stays
restricted to models that runtime can resolve.
Colocated regressions cover pin present on create and reopen, clearing
the pin returning a reopened branch to main's model and specifically not
the old pinned one, an unparseable pin behaving as no pin, and the
unknown-main-model fallback to no override.
* no-mistakes(review): Make unpinned supervision follow main model changes
* no-mistakes(document): Correct supervision model documentation
* feat(pi): let /supervision-model pick branch reasoning effort (#3079)
* feat(pi): let /supervision-model pick the branch's reasoning effort
Supervision is an easier job than the captain's own conversation, so the
Pi supervision branch does not need main's reasoning effort any more than
it needs main's model. /supervision-model now settles both in one flow:
the existing model picker, then a follow-up effort picker built from Pi's
own supported thinking levels for the model just chosen. Firstmate keeps
no effort catalog of its own; the menu, the clamp, and the vocabulary all
come from Pi.
The pick persists as one line in this home's gitignored
config/supervision-branch-effort, independent of the model pin: a captain
may pin a model, an effort, both, or neither. The effort pin's current
state decides the branch effort on every branch build - the first wake of
a cold start and the reopen after /new, /resume, /fork, or reload - and
overrides Pi's restore of whatever level a reopened branch session
recorded, which is what keeps "Follow main" honest. With no pin, main's
own current effort is applied explicitly and followed live through Pi's
thinking_level_select event, the same way an unpinned branch already
follows main's model, and the two selections now share one build revision
so either change invalidates an in-flight build.
The branch is never refused over effort. Pi owns the clamp, so a pinned
level the branch's model cannot run becomes that model's nearest supported
level while the captain's raw pick is kept for a model that supports it,
and the command reports the level the branch will really run at rather
than the raw pin. A token Pi would not recognize at all is treated as no
pin rather than passed to that clamp, which would otherwise collapse a
typo into the model's lowest level. Only when main's effort cannot be read
either does a build pass no effort override at all, which is the behavior
from before this file existed.
Pi's own effort vocabulary is pinned by a bidirectional type assertion
against Pi's getThinkingLevel return type, so the tracked strict typecheck
against the installed package fails the moment Pi adds or removes a level.
docs/configuration.md owns the operator-facing schema for both pins.
Portable regressions cover the pin on create and reopen, model-only and
effort-only pins working independently, clearing a pin returning the
branch to main's effort, live-follow of a mid-session change, the clamp,
an unrecognized token, the unknown-main-effort fallback, and the command's
two-step flow, persistence, cancellation, and honest reporting. The opt-in
real-SDK guard proves the vendor surface all of that rests on, and also
repairs a pre-existing gap that left it unable to load the extension at
all.
* no-mistakes(review): Resolve effective branch effort honestly
* no-mistakes(document): Clarify Pi-owned effort picker behavior
* fix: keep routine supervision noise out of captain chat (#3093)
* fix(supervision): silence empty board closes and decouple the heartbeat
Two unrelated sources of noise put routine supervision events in the
captain's chat.
An empty Lavish board close - the captain reads a review surface, says
nothing, and closes it - became a check wake whose entire content was
that nothing happened. Suppress it at its source instead of routing it
anywhere: the generic runner gains a `silent` adapter seam mirroring the
existing `terminal` one, and the Lavish adapter answers it for exactly
one positively-determined shape, an `ended` session carrying no queued
content block. A silenced result is recorded durably handled so it does
not return on a later reconcile. Everything else announces unchanged - a
`Send & End` close carrying the captain's real answer, an `ended` result
still carrying content, a waiting or missing session, an unreadable
result, and every adapter that implements no `silent` command at all.
The keyed-answer feed is untouched, so suppressing an announcement never
suppresses the captain's own answer.
A fleet heartbeat was deferred to main merely because some unrelated
check row happened to be sitting unread, which put a routine fleet
review in the chat for a reason that had nothing to do with the fleet. A
check row is permanently main-owned, so it is now excluded from a
heartbeat claim rather than vetoing the scan, exactly as in every other
mode. What all-or-nothing guarantees is unchanged: the branch takes
every branch-ownable unread row or none of them, and an unresolvable
task-local row, an unknown row kind, or an unreadable queue still defers
the whole review to main. Main is still woken for the check on that
check's own triggering close, so nothing starves.
Main-only classes are unchanged and now each covered by a test: Relay
mentions, credential failures, merge confirmations, real board answers,
and watcher-failure repair. The per-actor acknowledgement and
no-cross-swallow properties are untouched.
* no-mistakes(review): Fail closed on all Lavish content headers
* no-mistakes(review): Suppress false unacknowledged status for silenced results
* fix(bin): stop a correlation token from hiding and stranding decisions (#1967)
* fix(classify): read the decision fold through a correlation token
status_line_verb stripped a trailing [key=...] from a status line's prefix
but left everything else glued to the verb, so a line carrying the
correlation token bin/fm-pending-reply-lib.sh embeds and a secondmate
echoes back matched no arm of _fm_decision_fold_line. Such a line folded as
ordinary status in both directions: a needs-decision or blocked opener never
opened its key, and a resolved or captain-held closer never closed one. The
same glued verb also hid correlated done and blocked lines from
status_is_captain_relevant and status_is_terminal_verb, and let correlated
working and resolved lines leak through the free-text fallback the
nonterminal guard was meant to stop.
The verb parse now walks whole words and drops only a token of the exact
shape a firstmate library writes - corr=<16 hex>, plus the bracketed form
bin/fm-secondmate-report.sh emits - before or after the key token, unkeyed,
or doubled. An arbitrary name=value word is deliberately NOT skipped:
skipping unknown tokens would let free text carrying an equals sign reduce
to a bare verb and impersonate a transition, which is the takeover the
strict parse and _fm_decision_key_transition_allowed exist to prevent. A
prefix with no corr= substring is returned byte-for-byte as before, so every
line without a token keeps its exact historical verb.
FM_OPEN_DECISIONS_FOLD_VERSION goes to 3, because every cursor persisted
under the previous reading carries an open set computed while correlated
lines were invisible and must be rebuilt from byte 0.
Measured over a real 383-line status log: 254 lines keep byte-identical
captain-relevance, pause, terminal-verb and captain-held verdicts, and all
129 changed lines carry a valid token - 14 correlated done/blocked/
needs-decision lines become captain-visible, and 20 correlated
working/resolved lines stop being escalated on prose alone.
* fix(review): Captain, block token-first decision impersonation
* fix(document): Clarify normalized status verb ownership
* fix(classify): reconcile the correlation-token read with the tag-stop parser
Rebasing onto main put this change beside #2280, which made verb parsing
stop at the first "[name=value]" tag. Both edit status_line_verb with
different intents, so the resolution keeps both rules rather than letting
one overwrite the other:
- #2280's tag stop is kept verbatim and now owns every BRACKETED tag,
including the "[corr=...]" form fm-secondmate-report.sh writes. The
bracket-unwrapping arm this branch had added to the token test is
therefore removed as unreachable.
- This branch's token walk is kept and narrowed to the UNBRACKETED token
fm-pending-reply-lib.sh writes, which the tag stop does not reach.
Two consequences of standing beside #2280 rather than before it:
The fold version had collided at 4: #2280 spent it on the tag-stop
parser and this branch had spent it on the token read. A cursor
persisted under #2280's reading predates this one and must still be
rebuilt, so the version moves to 5.
A bracketed impostor is dropped from the malformed-token list. On main
today "resolved [corr=deadbeef] [key=victim]:" already reads as the bare
verb, as does "resolved [anything at all] [key=victim]:", because the tag
stop ends the parse at the opening bracket regardless of content. That is
#2280's reviewed contract; asserting otherwise here would narrow it. The
unbracketed impostors it owns stay strict and still fold as prose.
Adds a consumer test for the two verb-string case arms that postdate this
branch: fm-supervise-daemon.sh's transient-stale arm and
fm-crew-state.sh's map_log_state.
* fix(review): Captain: Seed cursor migration fixture with version four
* fix(document): Clarify voice status normalization ownership
* fix(bin): Cursor-Park unter Pi-Host ohne Cursor-Identität stilllegen (#3115)
* fix(bin): Cursor-Park unter Pi-Host stilllegen.
pi-cursor-sdk lädt .cursor/hooks.json in die Pi-Sitzung und parkte einen zweiten Watcher; das erzeugte rearm-resurface und brach laufende Rückfragen ab. Bei PI_CODING_AGENT=true beendet der Park sofort, native cursor-agent Primaries bleiben unverändert.
* fix(bin): Cursor-Park trotz PI-Leak nur ohne Cursor-Identität stilllegen.
Stand-down gilt nur bei PI_CODING_AGENT=true ohne CURSOR_AGENT und ohne CURSOR_INVOKED_AS. Handgestartete cursor-agent Primaries mit geerbtem PI-Marker parken weiter.
* no-mistakes(document): Document Cursor park Pi-host stand-down
* fix(bin): no-mistakes-Mindestversion auf 1.46.0 anheben.
Die PR-Attestierung verlangt ab 1.46.0 strukturierte Pipeline-Schritte; der Bootstrap-Floor blieb bei 1.31.2 und ließ zu alte Builds zu.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(pi): make supervision model picker searchable and scrollable (#3099)
* fix(pi): make /supervision-model's model list bounded and searchable
Pi's generic extension selector renders every option at once with no
search box, so a real eligible catalog ran off the top of the terminal.
The model step now draws the same rows through Pi's own SelectList - the
bounded scrolling primitive behind Pi's /model picker - with Pi's own
Input and fuzzy filter above it for search, keeping 'Follow main' first,
the branch-runtime eligibility filter intact, and the pick branch-only.
Pi's ModelSelectorComponent is deliberately not reused: its selection
handler writes the captain's default model through Pi's settings manager,
which would move main's conversation as a side effect of pinning the
branch. The effort step's menu is a handful of levels and stays on Pi's
plain selector dialog.
* no-mistakes(document): Clarify supervision picker documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(document): Document searchable supervision model picker
* no-mistakes: apply CI fixes
* fix(bin): durably report merged pull requests (#3104)
* fix: make a landed merge leave a durable outcome
A merge was the one lifecycle event that left no record outside the
merging agent's memory. bin/fm-pr-merge.sh ended at the forge call, and a
home merging under standing authority never waits for the merge poll that
would otherwise confirm it, so three real merges reached the captain as
silence.
bin/fm-merge-outcome-lib.sh is the single owner of that record. A
secondmate home reports the landed PR upward on the same parent reply
channel its terminal-outcome backstop already uses; a main home records
it on the durable wake queue. The record is at most once per task and
canonical PR identity, and only a merge that actually landed produces one.
The merge poll feeds that same channel when it detects a merge this home
did not perform, so the captain's own forge merge and a merge firstmate
performed itself produce one consistent outcome instead of two reporting
paths. No new state file and no second watch path.
Two smaller gaps from the same failure:
- A mate charter listed its report triggers without naming a landed
merge. Under standing merge authority nothing is ever "ready for
review", so the enumerated list silently omitted the case that matters.
- A secondmate home seeded without its parent binding failed every
terminal-outcome report for the same reason, and the diagnostic never
named the binding. It does now.
* no-mistakes(review): Harden durable merge outcome reporting
* no-mistakes(review): Make merge race regression deterministic
* no-mistakes(review): Make merge outcomes retry-idempotent and forge-confirmed
* no-mistakes(review): Unify merge publication under canonical outcome marker
* no-mistakes(review): Publish merge outcomes before committing dedup markers
* no-mistakes(review): Document at-least-once merge outcome recovery
* no-mistakes(review): Use supported GitHub confirmation and update recovery docs
* no-mistakes(review): Preserve distinct merge wakes by PR identity
* no-mistakes(document): Document durable merge outcome semantics
* no-mistakes(test): Make merge outcome interleaving test deterministic
* no-mistakes(document): Clarify merge outcome documentation ownership
* fix(lint): keep the merge-outcome library an analysis boundary
bin/fm-watch.sh followed the new merge-outcome library's source graph,
which reaches the wake queue, PR identity, and secondmate parent
libraries. Expanding that inside an already-large lint root pushed
ShellCheck's external-source analysis past the bounded CI lint worker:
the Lint job was killed with SIGTERM after five silent minutes, twice,
having emitted no diagnostics at all.
Make it an analysis boundary, exactly as the transition and inbox owners
directly above and below it already are and for the same stated reason.
Coverage is unchanged because the library is a canonical lint root in its
own right and is still linted as one.
Measured locally: the watcher goes from not terminating within 120s to
9s clean, and the library alone lints in 1s clean.
* fix(bearings): preserve projections through inventory mismatches (#3129)
* fix(bearings): keep an inventory-mismatch home readable, and mark warnings as repairs
A backlog-vs-metadata inventory mismatch inside a secondmate home was being
reported as "we cannot read that home", which discarded that home's open
captain calls, queued work, landed work, and live workers from the whole
Bearings digest. The main home already treats the identical mismatch as a
harmless disclosure; this makes the secondmate path agree.
- fm-fleet-snapshot.sh: the invalidity gate now passes orphan_in_flight,
unowned_current, and terminal_in_flight through the partial-structured
carve-out alongside child_current_unavailable, so those homes keep their
decisions, holds, queued, landed, and live work and leave unreadable[].
missing_backlog and unstructured_current stay on the discard path, because
there the backlog itself is untrustworthy.
- fm-fleet-snapshot.sh: the same three kinds no longer collapse the home's own
classification to "unknown"; the real captain_decision / active_child_work /
externally_held classification survives and invalidity carries the warning.
An unavailable child state still collapses it, including when a mismatch
masks it under strict-invalidity precedence.
- secondmate_landed.partial now keys on partial-structured trust rather than an
unknown state, so an inventory-mismatch home is still…
Bre77
added a commit
to Bre77/firstmate
that referenced
this pull request
Sep 2, 2026
…ts) (#72) * fix(ci): fail hung Herdr behavior runs in 20 minutes (#2413) A wedged family-run step was occupying the runner until the 75-minute job cap; bound that step so cleanup and timing artifacts still upload. * fix: keep the public promise reachable when work is routed to a second mate (#2457) The lightweight Relay follow-up link lives in the answering home's own state/<task-id>.meta, so it can only bind work that home owns. When a Relay-linked request is routed to a second mate, the task record lives in the second mate's home, fm-x-link.sh failed with a bare "no such task ...meta", and nothing else picked the promise up: only the soft acknowledgement was ever posted. The typed promised-final path already supports --work-home secondmate:<id>; the playbook simply never chose it. - fmx-respond now states the routing rule crisply: a task in this home takes the lightweight link, and second-mate-routed work takes a promised-final commitment bound to that home, registered up front with the brief command carried into the routed worker's instructions. - fm-x-link.sh refuses a task with no local record by naming the registered second mate whose home actually holds it and printing the promised-final registration command, with the exact --work-home when the match is unambiguous. A home with no registered second mates keeps the plain error. - fm-backlog-handoff.sh reports, after a successful move, any moved key that still owes a public reply bound to main/<key>, since that binding no longer names the home owning the work. The move itself is never blocked. Docs and the secondmate handoff prose follow the same rule. Tests cover the refusal, its scoping, the unchanged local-link path, and both handoff outcomes at the script boundary. * docs(skills): add remote-secondmate recovery hint for false-negative verdicts (#2456) * fix(skills): hint that remote secondmate liveness verdicts false-negative fm-crew-state and fm-send routinely misreport a live remote secondmate as dead; confirm against the pane before relaunching, and relaunch only through fm-spawn.sh, never raw herdr pane surgery. * no-mistakes: apply CI fixes * fix(calm): keep Pi's export confirmation visible (#2461) Pi 0.83.0 added a status line to every tool-expansion change, and Pi updates the previous status line in place when two status messages arrive back to back. Calm's post-export redraw cycled tool expansion on the macrotask right after Pi printed "Session exported to: <path>", so both expansion status lines coalesced over that confirmation and the captain was left with no record of where their export landed. Calm now repaints only the tool rows it presents, by invalidating each row through the render context Pi hands its render slots, and requests the surrounding redraw through setStatus. Neither appends to the transcript. The repaint is still needed because Pi can re-render a row asynchronously - the built-in edit row invalidates itself once its diff is ready - and that re-render can land inside the window where /export forces stock rendering. The real-terminal /export case now asserts the confirmation is still on screen after the redraw has settled, and that the redraw restored every Calm-hidden row, instead of only racing the moment the confirmation first appeared. * feat(stow): add open-record persistence to /stow before reset (#2488) * feat(stow): persist the open records a session is holding /stow curated memory and captured session knowledge, but never touched record state, while AGENTS.md called it an "unfinished-work sweep" and the receipt declared the session "safe to reset" - wording that implied a record-correctness guarantee stow does not make. A shipped PR with no backlog item, a queued umbrella whose phases had merged, and four decision holds left open after their answers shipped all survived repeated stows. Add a bounded pass that files record state from the same volatile input the rest of stow already uses: the open threads in context, minutes before the reset destroys them. It creates a record for an unfiled thread and corrects one the session knows is wrong, through the owning path, and states its boundary as part of the contract - it never enumerates the backlog, lists holds, or queries a forge, because it cannot be a reconciliation and must not be read as one. Correct the wording in AGENTS.md and the completion receipt so reset-safe means what it actually guarantees: nothing this session knew was lost. * no-mistakes(review): correct stow decision-hold inspection to read hold via tasks-axi * no-mistakes(document): note /stow open-record persistence in README command catalog * refactor(stow): state open-record persistence as principle, not procedure The first version enumerated triggers, named commands, and prescribed an ordered procedure. That is too rigid for an agent skill: it invites literal execution of a checklist instead of judgment, and every enumerated example is a way for the guidance to go stale. Reduce it to the intent - before a reset, the important open work you are holding in context must end up durably recorded rather than dying with the session, filing what is unfiled and correcting what is stale - and let the agent judge importance, the record, and the owning write path. Keep the scope bound, since it is a decided contract and not a mechanic: this covers the open work the session is holding, never a reconciliation of durable records against repository or forge reality. The wording corrections in AGENTS.md and the completion receipt are unchanged. * fix(decisions): close decision holds at answer time via one general keyed-answer path (#2490) * fix(decisions): close captain holds at answer time Firstmate had two "a decision is open" ledgers with asymmetric closing mechanics. The live status-log ledger closes atomically at answer time, because bin/fm-send.sh --resolve-key makes answering a decision be the act that closes it. The durable backlog hold ledger had no such coupling: answering and recording were two separate acts, and only the first was forced by the workflow. That asymmetry lost four real captain decisions. Their answers were captured durably to disk, keyed character for character by the hold decision keys, acknowledged, and even implemented and shipped, yet the holds stayed open for two days and the captain was asked to re-answer decisions already on his own disk. Give the hold ledger the same answer-time-closure property: - bin/fm-decision-hold.sh gains an `answer` subcommand, the hold ledger's counterpart to --resolve-key. It shares one unrouted close implementation with `decline`, so it carries every existing guard - the captain decision file, the active-hold requirement, retry identity, and the refusal to release still-routed work - and differs only in the resolution mode it records. `decline` keeps its stronger meaning that the answer routes no follow-up work at all. - bin/fm-procevent-lavish.sh wires the channel that actually carried the lost answers. `arm --decisions-origin` binds a deck to the origin whose holds it carries, `answers` reads the structured choices out of a captured poll result, `close-decisions` maps each key to its hold and closes it through the command above, and `autohandle` lets the runner apply that at capture time. Safety is preserved rather than traded away. Only rows tagged `choice` are read, so freeform captain prose cannot forge a decision key. Closure is confined to the one bound origin. The decision text is a pure function of the captured result, so a replayed capture is idempotent. A hold that is absent, already closed, or still blocking routed work is skipped and left for `resolve`, never forced. A deck armed without the binding touches no hold at all. And autohandle deliberately never reports full handling, because recording an answer is transcription while acting on it is firstmate's judgement - so the check wake still reaches the handler. fm-send --resolve-key is untouched. * no-mistakes(document): document state/lavish-decisions binding dir in AGENTS.md state inventory * refactor(decisions): make keyed-answer closure one general capability The previous pass gave holds answer-time closure but built it as bespoke Lavish wiring: the review adapter carried the source-to-origin binding, mapped keys to hold identities, wrote decision records, decided what to skip, and closed holds itself. That treated a review deck as a special decision source. It is not - it is an ephemeral discussion format that happens to carry answers. Collapse it into ONE general capability with one owner. bin/fm-decision-hold.sh now owns the whole of "a keyed answer closes its matching hold": - `answers <origin> --source <provenance>` is the channel-agnostic intake. It reads key/answer/label lines on stdin, maps each key to its hold, and closes it through the same `answer` path, so every guard applies identically whatever channel the answer came from. --source is provenance recorded in the decision, never a behavior switch; there is no per-channel branch and no knowledge of chat, decks, or transports. - `bind`/`unbind`/`binding` own the source-to-origin binding for any channel whose answers arrive detached from their origin. Every channel is now an ordinary caller that only turns what it received into keyed lines: - bin/fm-send.sh (chat) feeds the intake for a key that names an active hold. This also fixes a real gap: once `complete` transfers a decision to its hold it closes the live status copy, so --resolve-key alone could never answer a transferred decision. - bin/fm-procevent.sh feeds it generically. A bound source's captured result goes to `<adapter> answers <result-file>` and whatever that prints is piped into the intake. The runner names no adapter, parses no result, and carries no decision rule, so any future adapter with an `answers` command works with no change here. - bin/fm-procevent-lavish.sh keeps only `answers`, which reports the structured choices a review captured and stops. It maps nothing to a hold and closes nothing; it lost ~160 lines of decision logic. Feeding is independent of handling, so it never acknowledges a result and never suppresses a wake - recording an answer is transcription, acting on it stays firstmate's judgement. The regression that proves closure now drives a FIXTURE adapter that is not the review adapter, so what is proven is that any bound channel reaches the intake rather than that one channel is wired specially. A new regression drives the real fm-send over a stubbed transport for the chat side. Every prior guarantee still holds, and fm-send's status-log behavior is unchanged. * no-mistakes(review): test(decisions): drop source-content grep from hold-closure regression * fix(memory): emit a real @AGENTS.md pointer instead of a CLAUDE.md symlink (#2512) A Write aimed at CLAUDE.md followed the symlink and destroyed AGENTS.md. The installer now creates and migrates to a recoverable two-line pointer file. * fix(ci): keep CLAUDE.md pointer check valid (#2515) * ci: gate GitHub workflows with pinned actionlint (#2517) * fix(lint): catch malformed GitHub workflows before merge A self-broken ci.yml cannot report its own breakage, so parse every workflow in the local lint path that no-mistakes already runs. * fix(lint): pin actionlint instead of Ruby for workflow lint A self-broken ci.yml still has to fail in the local lint path, and the named tool for that gate is actionlint, not a new Ruby runtime. * no-mistakes(document): Clarify pinned workflow lint documentation * fix: install pinned lint tools across supported platforms (#2546) * fix: install pinned shellcheck and actionlint on macOS and linux arm64 The installers were hardcoded to linux amd64 and sha256sum, so a Mac dev could not satisfy the refuse-on-mismatch lint gate. Select the official per-platform archive and checksum, and fall back to shasum -a 256. * no-mistakes(document): Document cross-platform pinned lint installers * docs: reconcile test-evidence docs with store_in_repo: true (#2548) .no-mistakes.yaml has set test.evidence.store_in_repo: true since #2355, but CONTRIBUTING.md, docs/configuration.md, and docs/architecture.md still described the old policy of keeping evidence out of the repo in a temp directory. The current no-mistakes behavior for store_in_repo: true is to publish each run's test evidence to the orphan no-mistakes/evidence branch and link it from the PR body. That branch shares no history with code branches, so evidence never enters a pushed feature branch or the default branch, and CI's tracked personal fleet paths rule stays accurate. Docs only. No change to .no-mistakes.yaml or any workflow. * docs: clarify test evidence branch storage (#2549) * docs: correct test evidence storage comment in .no-mistakes.yaml * no-mistakes: apply CI fixes * docs: hint that live scouts may host their own Lavish review loop (#2563) Make that a first-class option in always-loaded instructions so firstmate does not default to mediating and tearing the scout down between iteration rounds. * fix(bin): report remote secondmate delivery and state truthfully (#2570) * fix(bin): report remote secondmate delivery and state truthfully A steer to a remote secondmate crosses fm-on.sh to a host-local fm-send leg whose unconfirmed submit read-back (verdict=pending, typically a busy mate whose harness queues the steer) was flattened into exit 1, so the parent printed "error: text not submitted" / "error: text not sent" and discarded the pending-reply expectation for a steer that had actually landed. fm-send now carries the verdict across the ssh boundary as a documented delivered-unconfirmed exit 3: the parent reports the steer as delivered with confirmation pending, exits 0, keeps the expectation armed (awaiting_report), and closes --resolve-key decisions, while transport loss (ssh 255) and real remote failures keep failing loudly with the remote leg's stderr attached. A local unconfirmed submit now also exits 3 with an honest non-error message and still never closes a decision key. fm-crew-state.sh and fm-peek.sh no longer read a remote mate's endpoint through local probes (which misreported a healthy mate as "worktree gone" / "can't find session: remote"): both now use the true remote source over fm-on.sh, and an unreachable or unreadable remote reads as unknown-remote, never as gone or dead. * no-mistakes(document): Document remote delivery and state truth * no-mistakes: apply CI fixes * feat: adopt spendPriority for quota dispatch (#2574) * Adopt quota-axi 0.1.29 spendPriority-primary array dispatch. quota-axi 0.1.29 publishes schema 5 with selection.spendPriority as the primary comparative signal and demotes derivation fields out of default --json. Rank comparable-fit candidates on that scalar, keep runway versus the completion horizon as a hard gate, and raise the compatibility floor so a pre-consolidation build cannot reach dispatch intake. * no-mistakes(review): Correct schema fixtures and remove prescriptive selection prompts * no-mistakes(document): Correct quota verification evidence chronology * Collapse quota-array-dispatch onto TOON-first spendPriority ranking. Decide from quota-axi's default TOON; keep --json as a rare defensive fallback. Rank by spendPriority after eligibility, reasoning-class, and runway-feasibility gates, and drop the hand-computed Pareto, pace, reserve, and window-id layers. * no-mistakes(review): Permit ambiguous JSON fallback and correct reset fixtures * no-mistakes(review): Correct runway semantics and escalate unresolved uncertainty * no-mistakes(document): Document TOON-first quota dispatch evidence * docs: add GROK_BOT.md Grok Bot system prompt (#2590) * docs: add GROK_BOT.md Grok Bot system prompt * docs: amend GROK_BOT.md with charter report-back and delegation marker * docs: classify GROK_BOT.md as public-product * docs: make GROK_BOT.md the plain Grok Bot system prompt * docs: update GROK_BOT.md nautical terms and self-improvement (#2592) * doc: Update language in GROK_BOT.md for clarity Refine language for clarity and consistency in instructions. * fix(bin): preserve inactive reconciliation scan progress (#2595) * fix(bin): guarantee inactive-reconcile scan progress under second quantization The inactive-outcome scan computed its aggregate deadline in whole seconds, so a 1-second budget's effective value lands anywhere in (0,1]; a scan starting just before a wall-clock second boundary rounded its whole budget away mid-scan and exited having visited no child, while the durable cursor had already advanced past the never-examined child. This is the CI flake behind tests/fm-inactive-reconcile.test.sh's 'next bounded scan did not resume with the following child' (watcher-wake-lock family, portable serial 2, seen on the PR #2590 run). Every scan now visits at least its first due child with the per-child state-read bound floored at one second, so no invocation can be a zero-work no-op. The outer process-group kill moves to budget+1s: the scan's own deadline enforces the budget, and the kill is a backstop for a scan wedged in an unbounded wait instead of a racer that routinely preempts the clean bounded exit. The wake-lock-wait test bound tracks the backstop (3s -> 4s); the previously flaky assertion is unchanged. * no-mistakes(document): Document inactive-reconcile deadline backstop * doc: Revise Firstmate delegation and communication guidelines Refactor the guidelines for Firstmate's role and delegation process, emphasizing the importance of crewmates and asynchronous work. * doc: Update work delegation and secret management instructions Clarified guidelines for handing off work to crewmates and managing secrets. * test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617) Three assertions in tests/fm-procevent.test.sh depended on a detached runner having finished work that the command starting it does not wait for. reconcile's replacement runner is started through detach_runner, which only forks: reconcile returns and counts the start before that runner has claimed its source or exec'd its child. Any assertion taken straight after reconcile therefore samples a race. - The publish-before-apply recovery section left its always-ready /bin/echo source registered across the recovery reconcile, so that reconcile launched a competing detached poll (observed: started=1) that then raced every later assertion for the source claim, the next capture sequence, and this home's applied record, and outlived the section holding a live claim. It is now retired before that reconcile - re-announcement is proven from the durable inbox alone and needs no registration - and started=0 is asserted so a competing poll cannot be reintroduced unnoticed. This is the same retire-before-reconcile discipline the self-announcing section already carries; that section acquired it after the identical race made its "not-autohandled: self-src" assertion read "already owned: self-src". - The crashed-leader replacement section snapshotted the replacement's claim file and execution log behind a fixed 0.5s settle window. On a loaded machine that window expires first, which is the CI flake behind "a replacement runner started without recording its own claim" and "reconcile did not start exactly one replacement source". Both effects are now waited for with the suite's bounded wait helpers; the exact one-replacement count is still asserted afterwards, unchanged. - The duplicate-start section slept 0.5s for reconcile's runner to record ownership before asserting that a second start loses to it. It now waits for that claim. Also tighten one assertion that could not fail as written: "autohandled: self-src" is a substring of "not-autohandled: self-src", so the applied path was accepted even when the runner reported the capture left for the handler. Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced 6 failing runs, all in the crashed-leader section. On the fixed suite, 216 full runs under the same load produced none. Reverting the self-announcing section's retire-before-reconcile line reproduces "already owned: self-src" on the first iteration, confirming the shared mechanism. * fix: preserve pending replies and defer remote reposts (#2618) * fix(bin): keep pending-reply expectations honest on both send legs Two related asymmetries let the parent-owned secondmate reply guard drop or nag requests it should not have. Local delivered-unconfirmed dropped the expectation. A marked request whose submit read-back stayed unconfirmed (verdict=pending) is the same not-a-failure outcome the remote leg reports as delivered, but fm-send discarded the parent's pending-reply record for it, so a request that very likely landed stopped being tracked entirely. The record now stays armed on its unconfirmed-delivery marker: a correlated report still resolves it, and an unanswered one still surfaces through the library's own reconciliation. Exit 3 and the local rule that an unconfirmed answer never closes a decision key are unchanged. Remote replies were nagged for a repost they did not need. A remote mate's report reaches the parent's status log only through the asynchronous mirror in fm-procevent-remote-reply.sh, yet the guard read an absent correlated line as proof the mate never reported - even while the answer was still in flight, which is the common case because the mirror's poll window is comparable to the recovery grace. The mirror now publishes one caught-up watermark from a quiet window, and the guard admits a missing report as evidence only once that watermark passes the turn that should have produced it. A genuinely missed report still gets exactly one repost, and a channel that is behind, unarmed, or broken leaves the request durably open and un-nagged rather than nagging blind; the mirror escalates its own continuity failures as before. Tests: a local unconfirmed secondmate send keeps its expectation armed and resolvable; a mirrored correlated remote reply resolves with no repost; a stale or absent watermark withholds the repost while a fresh one still releases it; a quiet remote window publishes the watermark and retirement clears it. * no-mistakes(review): Distinguish preempted polls from quiet windows * no-mistakes(document): Clarify remote reply channel freshness * no-mistakes(lint): Annotate shared remote preemption exit constant * fix(bin): honor declared pauses in busy-pane wedge checks (#2619) * fix(watch): honor a declared pause on a busy pane's completed-turn bound A worker that declares an external wait (`paused:`) and then blocks in one long foreground call - a review-hosting scout parked in a single blocking `lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane BUSY, so the stale path that already honors declared pauses never ran for it. The busy-pane completed-turn bound instead routed it straight into wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long as the review stayed open. busy_turn_bound_check now owns which absorber takes a crossed bound: a crew whose own last status line declares an external wait or a verified captain-held transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else keeps the unchanged wedge timer. The discriminator is the declaration together with liveness (the caller has already confirmed the pane is busy), never a blanket silencing - a crew that declared nothing, or whose pane is not live, escalates exactly as before, and a declared pause still re-surfaces once per long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched: the daemon owns pause triage there and already reads the same vocabulary. The two call sites also no longer clear pause bookkeeping in the same poll the pause cadence recorded it, which would have erased the re-surface throttle and turned the long cadence back into a per-poll re-surface. Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence recheck, and the restored wedge escalation once the declaration is lifted on the same busy over-age pane. Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded machine: fixed liveness budgets were reaping watchers mid-startup, so assertions on post-poll state passed vacuously or failed spuriously. Waits that describe a poll's outcome now wait for a completed poll cycle via the liveness beacon, the heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit budget is the uniform 10s already used elsewhere in the file. * no-mistakes(review): Fail poll-cycle waits on timeout * no-mistakes(review): Prevent poll timeout test hangs * no-mistakes(document): Clarify paused busy-pane supervision * doc: Enhance communication guidelines for decision-making Added guidelines for decision communication to the captain. * doc: Update task delegation and communication guidelines Clarify communication protocols with crewmates regarding task delegation and reporting. * fix(bin): reliably confirm herdr steer submission (#2647) * fix(herdr): confirm local steers that native agent-state misses Herdr can leave agent_status idle for a landed Claude turn and can keep queued Enter text visible while busy, so fm-send was reporting false swallows. Confirm those cases through the shared queued-Enter verdict and a cleared composer, and keep a genuine idle pending composer as unconfirmed. * no-mistakes(review): Stop Herdr Enter retries on unreadable composers * no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail * no-mistakes(review): Prevent confirmation after failed Herdr Enter * no-mistakes(review): Pace Herdr retries and clarify submit fallback * no-mistakes(review): Align Herdr submit docs with idle fallback * no-mistakes(document): Correct Herdr submit-confirmation documentation * feat(bearings): add interactive Lavish fleet board (#2659) * feat(bin): accept any-origin decision bindings with full-identity keys An aggregation surface (the bearings board) carries captain answers for holds across origins, but a binding was one-origin-per-source and the Lavish adapter capped question keys at 64 chars while real full hold identities measure 69-81. - fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker; binding prints it verbatim and answers accepts it, so the runner's feed seam carries an any-origin source with no runner change. In any-origin mode each key is a full hold identity <origin>-decision-<key>, split at its first -decision-; a key with no separator (merge/dispatch instructions) is skipped and feeds nothing, keeping non-decision answers out of the hold ledger by construction. Every existing close guard applies unchanged. - fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold identity fits; the slug-shape security property is unchanged. - tests: cross-origin closure through the real runner seam, an 81-char identity through the adapter, cap and shape refusals, routed-work skips, nonexistent-identity skips, and idempotent replay. * feat(bearings): add the /bearings lavish interactive fleet board /bearings lavish renders the bearings snapshot onto a shipped, reusable board template and arms it as a Lavish process-event source, so the captain answers Captain's Call items on the board and firstmate is woken by an ordinary check wake - no conversational turn ever blocks on a poll. - .agents/skills/bearings/assets/board-template.html: the shipped template (myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot, fail-closed schema guard that renders an error card instead of an empty fleet). Per-invocation agent work is composing the payload only. - bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload validation, slot injection with a round-trip check and \u003c escaping, stable board path, any-origin bind ALWAYS before arm, arm-if-absent. - bearings SKILL.md: the lavish invocation option, board composition rules, board-wake handling, and the captain-ruled merge-click authorization with its mandatory safeguards (PR resolved from the task's own meta record, wake-time green re-verification, never a red or changed PR, merges only through bin/fm-pr-merge.sh, chat echo with the full PR URL). - process-event-sources SKILL.md: one-line board-wake routing trigger. - tests: payload refusals, injection round-trip, bind-before-arm, idempotent re-arm, and template slot integrity. Fleet pickup: homes receive this after merge plus a firstmate self-update; landing timing is coordinated with the main firstmate. * no-mistakes(review): Harden bearings board validation and wake handling * no-mistakes(review): Require HTTPS for bearings board PR links * no-mistakes(review): Fail closed and bound bearings board answers * no-mistakes(review): Enforce UTF-8 byte limits for board answers * no-mistakes(review): Serve bearings board before arming and reject empty actions * no-mistakes(review): Prove bind-before-arm ordering through live answer consumption * no-mistakes(document): Document bearings board and cross-origin answers * fix(bearings): restore decision options and add close controls (#2707) * fix(bearings): always show decision options and a close/drop control Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call. * no-mistakes(review): Fix drop closure and decision-only option validation * no-mistakes(review): Preserve answerability for non-decision cards * no-mistakes(document): Clarify decision drop documentation * ci: require no-mistakes pipeline step attestation (#2710) Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed. * feat: collapse decisions into tasks held for the captain (#2728) * feat(captain-hold): collapse the decisions concept into tasks held for the captain A decision is no longer a separate type: it is an ordinary backlog task held for the captain, identified by its task id. bin/fm-captain-hold.sh owns the surviving behaviors - guarded hold creation, the recorded-answer close (answer/answers with a release mode for captain-gated work), the source bindings, and the investigation completion gate - and bin/fm-decision-hold.sh becomes a one-release compatibility shim over it. The fleet snapshot now parses hold-until and computes captain_actionable as queued + captain-held + unblocked + due, independent of row kind, plus a presentation-only deferred_marker for prose-deferred rows. Bearings renders every due captain-held task in Captain's Call, date-deferred holds as dated Charted Next gates, suppresses prose-deferred rows from default views with an omitted disclosure, and excludes from Recently Landed anything that closed while still held for the captain. Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already plain task ids and keep working; short keys in recorded metadata, concrete origin bindings, chat --resolve-key fallbacks, and old resolution records all resolve in place. * no-mistakes(review): Fix captain answer replay and body preservation * no-mistakes(review): Fix captain hold idempotency and legacy replay * no-mistakes(review): Validate card close modes and compatibility routing * no-mistakes(review): Enforce release replay mode matching * no-mistakes(review): Prevent duplicate decision cards and released replay mismatches * no-mistakes(review): Preserve answer columns and legacy resolve replays * no-mistakes(document): Document strict replay and legacy compatibility * no-mistakes(lint): Quote done literals to satisfy ShellCheck * no-mistakes: apply CI fixes * fix(rebase): keep collapsed captain hold board semantics * fix: bound recovery announcements and preserve supervision (#2733) * fix(watch): announce recovery once per generation and keep successors supervising A lost Pi/OpenCode handling handshake re-announced the same recovery generation on every cycle and spent the successor's first ~55s blind, so a real crew event could be ignored and then dropped. Record the announcement in the durable marker, confirm the handshake before the follow-up without swallowing failure, and enter the poll loop immediately. * no-mistakes(review): Tighten recovery event timing regression * no-mistakes(document): Document recovery-loop supervision guarantees * fix(bin): surface captain-call record divergence (#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue #2251's PR #2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes afte…
adonis-garcia-git
added a commit
to adonis-garcia-git/firstmate
that referenced
this pull request
Sep 3, 2026
…iation (#12) * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after #2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (#2857) * feat(bin): deliver local steers through durable task inboxes (#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes #2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (#2939) * fix(bin): bound remote job worker supervisor restarts (#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix: safely split supervision wake handling by actor (#2953) * feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup Three related fixes to the shared wake-drain and Pi supervision-branch dispatch machinery so a routine success is never main-blocking and a mixed queue can safely split between actors. 1. Successful routine results no longer create main-blocking wake rows. fm-startup-network.sh only enqueues a check: startup-network wake when the deferred result is actionable (state is not "done", or the report carries a bootstrap-diagnostics actionable prefix); a clean success stays durable in the report file without ever waking the agent. 2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes presentation and --ack-through to the current actor (bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original whole-queue cutoff behavior, unaffected. A branch actor (FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision branch's own bash tool calls) is scoped to an explicit eligible-row snapshot instead of a cutoff comparison, so it can never remove a row it was not granted - the fix for the swallow risk that used to force an all-or-nothing whole-queue fallback to main. .pi/extensions/lib/fm-branch-dispatch.ts's scopeForUnreadWake is the single owner of eligibility: a check-kind row (merge-confirmation polls, Relay mentions, credential/auth failures) is now excluded rather than vetoing the whole scan for a non-heartbeat wake, while a heartbeat review keeps its original all-or-nothing rule unchanged. writeEligibleRowsSnapshot publishes the exact eligible sequence numbers before every branch prompt; fm-primary-pi-watch.ts's offer still refuses a check-kind trigger outright so a main-only close is never itself routed to the branch. 3. A repeat identical merged-PR-poll result for an already-notified task is absorbed instead of enqueued again. A poll's own retirement state is scoped to one registration and cannot see a prior registration's outcome, so a task re-registered after its merge was already surfaced would otherwise wake main a second time for the same event. bin/fm-pr-lib.sh's new per-task pr-poll-merge-notified marker survives across re-registrations to catch that case; the first notification for a task still reaches main unchanged. Regression tests colocated in tests/fm-startup-network.test.sh, tests/fm-wake-queue.test.sh (including the mixed-queue no-swallow property), tests/fm-pi-branch-extension.test.sh, and tests/fm-pr-check-security.test.sh. docs/watcher-continuity.md and docs/pi-supervision-branch.md updated for the new contracts. * no-mistakes(review): Bind merge deduplication to canonical PR identity * no-mistakes(review): Serialize wake row ownership across main and branch * no-mistakes(review): Bind branch grants and deduplicate within actor claims * no-mistakes(review): Fallback main-owned wake claims to main delivery * no-mistakes(review): Clarify silent startup success guidance * no-mistakes(review): Release residual branch grants after settled prompts * no-mistakes(review): Reject truncated wake rows as corrupted * no-mistakes(document): Document per-actor routing and silent startup success * no-mistakes(lint): Fix ShellCheck findings in wake grant and startup test * no-mistakes: apply CI fixes * fix(pi): hide branch outcomes tool rows in Calm (#3024) * Hide branch outcome tool in Pi Calm * no-mistakes(review): Preserve stock outcomes rendering and document tool audit * no-mistakes(review): Document branch read tool audit disposition * no-mistakes(review): Match stock outcomes output sanitization * no-mistakes(document): Document Calm custom-tool visibility * fix: bind no-mistakes attestations to PR head (#3027) * fix: delegate no-mistakes PR gate to pinned action * no-mistakes(document): Document commit-bound no-mistakes attestations * feat(pi): add persistent supervision branch model selection (#3028) * feat(pi): let operators pin a cheaper supervision-branch model Supervision is an easier job than the captain's own conversation, so the Pi supervision branch does not need main's model. A new /supervision-model command opens Pi's own selector over Pi's own catalog of credentialed models, plus a "Follow main" entry, and persists the pick as one <provider>/<model-id> line in this home's gitignored config/supervision-branch-model. Firstmate keeps no model catalog of its own. The branch resolves the pin at every branch build - the first wake of a cold start and the reopen after /new, /resume, /fork, or reload - so the choice survives all of them, and picking also releases the live branch so the next wake reopens the same persistent branch conversation under the new model. An absent, unreadable, or unparseable file means no pin and keeps today's behavior byte for byte: no model option is passed and Pi picks the branch's model exactly as before. A pin naming a model Pi cannot hand back is never silently downgraded onto main's model: the branch refuses to build and the wake falls back to the captain-facing main path naming the unusable pin, which is the extension's existing failure direction. The choice is home-local and not part of secondmate inherited configuration, matching the Pi Calm preference precedent. docs/configuration.md owns the operator-facing schema. Portable regressions cover pin-present on create and reopen, pin-absent default, the command's persistence, cancellation, and live rebind, and both unusable and unparseable pins. The opt-in real-SDK guard proves the vendor surface the pin reads and that an explicit model wins over the model a reopened session recorded. * no-mistakes(review): Fix supervision model runtime and rebind races * no-mistakes(review): Restrict supervision picker to isolated runtime models * no-mistakes(document): Document supervision branch model selection * fix(pi): make the supervision model pin authoritative on every reopen Clearing the pin with "Follow main" removed the file but the next branch build reopened the persistent branch session with no explicit model override, so Pi restored the model that session had recorded - the old pinned model - while the command reported that the branch now follows main. The same gap meant an absent pin did not reliably mean same-model-as-main once a home had pinned once. The pin file's current state now decides the model on every branch build, create and reopen alike, overriding Pi's session-state restore. With a pin, that model. With no pin, main's own current model is applied explicitly, tracked from the contexts Pi already hands the extension plus its model_select event, since the branch is built at wake time with no context of its own. Only when main's model is unknown, or this home's stored credentials cannot run it in the isolated branch runtime, does a build fall back to passing no override at all, which is the behavior from before the pin existed; the branch is never refused over model choice. The command's notification now reports the model actually applied, and says plainly when clearing the pin could not apply main's model instead of claiming a change that did not take effect. No credential handling changes: the branch still relies entirely on the stored credentials its own runtime already holds, and the picker stays restricted to models that runtime can resolve. Colocated regressions cover pin present on create and reopen, clearing the pin returning a reopened branch to main's model and specifically not the old pinned one, an unparseable pin behaving as no pin, and the unknown-main-model fallback to no override. * no-mistakes(review): Make unpinned supervision follow main model changes * no-mistakes(document): Correct supervision model documentation * feat(pi): let /supervision-model pick branch reasoning effort (#3079) * feat(pi): let /supervision-model pick the branch's reasoning effort Supervision is an easier job than the captain's own conversation, so the Pi supervision branch does not need main's reasoning effort any more than it needs main's model. /supervision-model now settles both in one flow: the existing model picker, then a follow-up effort picker built from Pi's own supported thinking levels for the model just chosen. Firstmate keeps no effort catalog of its own; the menu, the clamp, and the vocabulary all come from Pi. The pick persists as one line in this home's gitignored config/supervision-branch-effort, independent of the model pin: a captain may pin a model, an effort, both, or neither. The effort pin's current state decides the branch effort on every branch build - the first wake of a cold start and the reopen after /new, /resume, /fork, or reload - and overrides Pi's restore of whatever level a reopened branch session recorded, which is what keeps "Follow main" honest. With no pin, main's own current effort is applied explicitly and followed live through Pi's thinking_level_select event, the same way an unpinned branch already follows main's model, and the two selections now share one build revision so either change invalidates an in-flight build. The branch is never refused over effort. Pi owns the clamp, so a pinned level the branch's model cannot run becomes that model's nearest supported level while the captain's raw pick is kept for a model that supports it, and the command reports the level the branch will really run at rather than the raw pin. A token Pi would not recognize at all is treated as no pin rather than passed to that clamp, which would otherwise collapse a typo into the model's lowest level. Only when main's effort cannot be read either does a build pass no effort override at all, which is the behavior from before this file existed. Pi's own effort vocabulary is pinned by a bidirectional type assertion against Pi's getThinkingLevel return type, so the tracked strict typecheck against the installed package fails the moment Pi adds or removes a level. docs/configuration.md owns the operator-facing schema for both pins. Portable regressions cover the pin on create and reopen, model-only and effort-only pins working independently, clearing a pin returning the branch to main's effort, live-follow of a mid-session change, the clamp, an unrecognized token, the unknown-main-effort fallback, and the command's two-step flow, persistence, cancellation, and honest reporting. The opt-in real-SDK guard proves the vendor surface all of that rests on, and also repairs a pre-existing gap that left it unable to load the extension at all. * no-mistakes(review): Resolve effective branch effort honestly * no-mistakes(document): Clarify Pi-owned effort picker behavior * fix: keep routine supervision noise out of captain chat (#3093) * fix(supervision): silence empty board closes and decouple the heartbeat Two unrelated sources of noise put routine supervision events in the captain's chat. An empty Lavish board close - the captain reads a review surface, says nothing, and closes it - became a check wake whose entire content was that nothing happened. Suppress it at its source instead of routing it anywhere: the generic runner gains a `silent` adapter seam mirroring the existing `terminal` one, and the Lavish adapter answers it for exactly one positively-determined shape, an `ended` session carrying no queued content block. A silenced result is recorded durably handled so it does not return on a later reconcile. Everything else announces unchanged - a `Send & End` close carrying the captain's real answer, an `ended` result still carrying content, a waiting or missing session, an unreadable result, and every adapter that implements no `silent` command at all. The keyed-answer feed is untouched, so suppressing an announcement never suppresses the captain's own answer. A fleet heartbeat was deferred to main merely because some unrelated check row happened to be sitting unread, which put a routine fleet review in the chat for a reason that had nothing to do with the fleet. A check row is permanently main-owned, so it is now excluded from a heartbeat claim rather than vetoing the scan, exactly as in every other mode. What all-or-nothing guarantees is unchanged: the branch takes every branch-ownable unread row or none of them, and an unresolvable task-local row, an unknown row kind, or an unreadable queue still defers the whole review to main. Main is still woken for the check on that check's own triggering close, so nothing starves. Main-only classes are unchanged and now each covered by a test: Relay mentions, credential failures, merge confirmations, real board answers, and watcher-failure repair. The per-actor acknowledgement and no-cross-swallow properties are untouched. * no-mistakes(review): Fail closed on all Lavish content headers * no-mistakes(review): Suppress false unacknowledged status for silenced results * fix(bin): stop a correlation token from hiding and stranding decisions (#1967) * fix(classify): read the decision fold through a correlation token status_line_verb stripped a trailing [key=...] from a status line's prefix but left everything else glued to the verb, so a line carrying the correlation token bin/fm-pending-reply-lib.sh embeds and a secondmate echoes back matched no arm of _fm_decision_fold_line. Such a line folded as ordinary status in both directions: a needs-decision or blocked opener never opened its key, and a resolved or captain-held closer never closed one. The same glued verb also hid correlated done and blocked lines from status_is_captain_relevant and status_is_terminal_verb, and let correlated working and resolved lines leak through the free-text fallback the nonterminal guard was meant to stop. The verb parse now walks whole words and drops only a token of the exact shape a firstmate library writes - corr=<16 hex>, plus the bracketed form bin/fm-secondmate-report.sh emits - before or after the key token, unkeyed, or doubled. An arbitrary name=value word is deliberately NOT skipped: skipping unknown tokens would let free text carrying an equals sign reduce to a bare verb and impersonate a transition, which is the takeover the strict parse and _fm_decision_key_transition_allowed exist to prevent. A prefix with no corr= substring is returned byte-for-byte as before, so every line without a token keeps its exact historical verb. FM_OPEN_DECISIONS_FOLD_VERSION goes to 3, because every cursor persisted under the previous reading carries an open set computed while correlated lines were invisible and must be rebuilt from byte 0. Measured over a real 383-line status log: 254 lines keep byte-identical captain-relevance, pause, terminal-verb and captain-held verdicts, and all 129 changed lines carry a valid token - 14 correlated done/blocked/ needs-decision lines become captain-visible, and 20 correlated working/resolved lines stop being escalated on prose alone. * fix(review): Captain, block token-first decision impersonation * fix(document): Clarify normalized status verb ownership * fix(classify): reconcile the correlation-token read with the tag-stop parser Rebasing onto main put this change beside #2280, which made verb parsing stop at the first "[name=value]" tag. Both edit status_line_verb with different intents, so the resolution keeps both rules rather than letting one overwrite the other: - #2280's tag stop is kept verbatim and now owns every BRACKETED tag, including the "[corr=...]" form fm-secondmate-report.sh writes. The bracket-unwrapping arm this branch had added to the token test is therefore removed as unreachable. - This branch's token walk is kept and narrowed to the UNBRACKETED token fm-pending-reply-lib.sh writes, which the tag stop does not reach. Two consequences of standing beside #2280 rather than before it: The fold version had collided at 4: #2280 spent it on the tag-stop parser and this branch had spent it on the token read. A cursor persisted under #2280's reading predates this one and must still be rebuilt, so the version moves to 5. A bracketed impostor is dropped from the malformed-token list. On main today "resolved [corr=deadbeef] [key=victim]:" already reads as the bare verb, as does "resolved [anything at all] [key=victim]:", because the tag stop ends the parse at the opening bracket regardless of content. That is #2280's reviewed contract; asserting otherwise here would narrow it. The unbracketed impostors it owns stay strict and still fold as prose. Adds a consumer test for the two verb-string case arms that postdate this branch: fm-supervise-daemon.sh's transient-stale arm and fm-crew-state.sh's map_log_state. * fix(review): Captain: Seed cursor migration fixture with version four * fix(document): Clarify voice status normalization ownership * fix(bin): Cursor-Park unter Pi-Host ohne Cursor-Identität stilllegen (#3115) * fix(bin): Cursor-Park unter Pi-Host stilllegen. pi-cursor-sdk lädt .cursor/hooks.json in die Pi-Sitzung und parkte einen zweiten Watcher; das erzeugte rearm-resurface und brach laufende Rückfragen ab. Bei PI_CODING_AGENT=true beendet der Park sofort, native cursor-agent Primaries bleiben unverändert. * fix(bin): Cursor-Park trotz PI-Leak nur ohne Cursor-Identität stilllegen. Stand-down gilt nur bei PI_CODING_AGENT=true ohne CURSOR_AGENT und ohne CURSOR_INVOKED_AS. Handgestartete cursor-agent Primaries mit geerbtem PI-Marker parken weiter. * no-mistakes(document): Document Cursor park Pi-host stand-down * fix(bin): no-mistakes-Mindestversion auf 1.46.0 anheben. Die PR-Attestierung verlangt ab 1.46.0 strukturierte Pipeline-Schritte; der Bootstrap-Floor blieb bei 1.31.2 und ließ zu alte Builds zu. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> * fix(pi): make supervision model picker searchable and scrollable (#3099) * fix(pi): make /supervision-model's model list bounded and searchable Pi's generic extension selector renders every option at once with no search box, so a real eligible catalog ran off the top of the terminal. The model step now draws the same rows through Pi's own SelectList - the bounded scrolling primitive behind Pi's /model picker - with Pi's own Input and fuzzy filter above it for search, keeping 'Follow main' first, the branch-runtime eligibility filter intact, and the pick branch-only. Pi's ModelSelectorComponent is deliberately not reused: its selection handler writes the captain's default model through Pi's settings manager, which would move main's conversation as a side effect of pinning the branch. The effort step's menu is a handful of levels and stays on Pi's plain selector dialog. * no-mistakes(document): Clarify supervision picker documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(document): Document searchable supervision model picker * no-mistakes: apply CI fixes * fix(bin): durably report merged pull requests (#3104) * fix: make a landed merge leave a durable outcome A merge was the one lifecycle event that left no record outside the merging agent's memory. bin/fm-pr-merge.sh ended at the forge call, and a home merging under standing authority never waits for the merge poll that would otherwise confirm it, so three real merges reached the captain as silence. bin/fm-merge-outcome-lib.sh is the single owner of that record. A secondmate home reports the landed PR upward on the same parent reply channel its terminal-outcome backstop already uses; a main home records it on the durable wake queue. The record is at most once per task and canonical PR identity, and only a merge that actually landed produces one. The merge poll feeds that same channel when it detects a merge this home did not perform, so the captain's own forge merge and a merge firstmate performed itself produce one consistent outcome instead of two reporting paths. No new state file and no second watch path. Two smaller gaps from the same failure: - A mate charter listed its report triggers without naming a landed merge. Under standing merge authority nothing is ever "ready for review", so the enumerated list silently omitted the case that matters. - A secondmate home seeded without its parent binding failed every terminal-outcome report for the same reason, and the diagnostic never named the binding. It does now. * no-mistakes(review): Harden durable merge outcome reporting * no-mistakes(review): Make merge race regression deterministic * no-mistakes(review): Make merge outcomes retry-idempotent and forge-confirmed * no-mistakes(review): Unify merge publication under canonical outcome marker * no-mistakes(review): Publish merge outcomes before committing dedup markers * no-mistakes(review): Document at-least-once merge outcome recovery * no-mistakes(review): Use supported GitHub confirmation and update recovery docs * no-mistakes(review): Preserve distinct merge wakes by PR identity * no-mistakes(document): Document durable merge outcome semantics * no-mistakes(test): Make merge outcome interleaving test deterministic * no-mistakes(document): Clarify merge outcome documentation ownership * fix(lint): keep the merge-outcome library an analysis boundary bin/fm-watch.sh followed the new merge-outcome library's source graph, which reaches the wake queue, PR identity, and secondmate parent libraries. Expanding that inside an already-large lint root pushed ShellCheck's external-source analysis past the bounded CI lint worker: the Lint job was killed with SIGTERM after five silent minutes, twice, having emitted no diagnostics at all. Make it an analysis boundary, exactly as the transition and inbox owners directly above and below it already are and for the same stated reason. Coverage is unchanged because the library is a canonical lint root in its own right and is still linted as one. Measured locally: the watcher goes from not terminating within 120s to 9s clean, and the library alone lints in 1s clean. * fix(bearings): preserve projections through inventory mismatches (#3129) * fix(bearings): keep an inventory-mismatch home readable, and mark warnings as repairs A backlog-vs-metadata inventory mismatch inside a secondmate home was being reported as "we cannot read that home", which discarded that home's open captain calls, queued work, landed work, and live workers from the whole Bearings digest. The main home already treats the identical mismatch as a harmless disclosure; this makes the secondmate path agree. - fm-fleet-snapshot.sh: the invalidity gate now passes orphan_in_flight, unowned_current, and terminal_in_flight through the partial-structured carve-out alongside child_current_unavailable, so those homes keep their decisions, holds, queued, landed, and live work and leave unreadable[]. missing_backlog and unstructured_current stay on the discard path, because there the backlog itself is untrustworthy. - fm-fleet-snapshot.sh: the same three kinds no longer collapse the home's own classification to "unknown"; the real captain_decision / active_child_work / externally_held classification survives and invalidity carries the warning. An unavailable child state still collapses it, including when a mismatch masks it under strict-invalidity precedence. - secondmate_landed.partial now keys on partial-structured trust rather than an unknown state, so an inventory-mismatch home is still disclosed as partial. Ask the home that owns the wrong books to fix them: - bin/fm-secondmate-reconcile.sh sends exactly one reconcile instruction per mismatch episode through the ordinary steering transport. A persistent mismatch keeps its episode identity and never re-nags; a changed mismatch earns one more ask; a repaired one is forgotten so a recurrence is asked about again. The parent never touches the mate's own files, and a failed send records nothing so the next run retries it. Give integrity warnings their own look on the board: - charted rows take an optional kind of "queued" (the default) or "warning". A warning badges "needs repair" instead of "waiting" and is excluded from the Charted Next count, so alarms stop reading as dispatchable queued work. No fifth board section, and every existing payload stays valid. Tests pin the new policy behaviorally: the retained surfaces and classification for all three mismatch kinds, the still-discarding unstructured_current and missing_backlog cases, the once-per-episode reconcile ask through real durable steering records, and the board rendering exercised through the shipped template under a minimal DOM shim. * no-mistakes(review): Make reconcile dedupe atomic and warnings non-dispatchable * no-mistakes(review): Preserve reconcile identity and reject stale snapshots * no-mistakes(review): Order snapshots uniquely and canonicalize episode identities * no-mistakes(review): Add fire-and-forget reconcile and separate warning overflow * no-mistakes(review): Exclude fire-and-forget from escalation and track reconcile background * no-mistakes(review): Run reconcile enqueue inline across all adapters * no-mistakes(review): Track reconcile clears across strict-invalidity homes * no-mistakes(review): Persist reconcile transitions atomically * no-mistakes(document): Document reconcile and fire-and-forget contracts * refactor(bearings): replace the reconcile episode dedupe with a 4-hour cooldown The reconcile ask needed to fire once per problem without nagging on every recap. The episode-precise record that tried to do that had to be correct in every direction at once - order two concurrent snapshots, tell a repair from a new problem, and never lose a clear - and each direction it got wrong either swallowed a nudge or sent a duplicate. A per-home cooldown removes the whole class. One durable timestamp per home, one nudge per four hours, and nothing to get stale, mis-order, or mis-classify: a home in mismatch is asked once, later recaps stay silent, and a mismatch still sitting there after the window earns one gentle re-nudge. - bin/fm-secondmate-reconcile.sh: state/<id>.reconcile-nudged holds the epoch second of the last ask; FM_RECONCILE_COOLDOWN_SECONDS names the window. The episode identity, ordering generation, pending/clear transitions, and delivery-identity reuse are all gone. A known-undelivered send starts no cooldown so the next run retries it; an unconfirmed one does, because a duplicate ask is worse than one the mate may already hold. - bin/fm-fleet-snapshot.sh, bin/fm-bearings-snapshot.sh: drop the snapshot `observation` monotonic identity, which existed only to order those records. - bin/fm-teardown.sh: retire the cooldown record with the endpoint's other runtime artifacts, so reseeding a retired id is not silenced by its predecessor's window. The inline durable fire-and-forget send is unchanged, and the projection fix and the warning surface are untouched. Tests follow the behavior: the cooldown suite now pins one ask per window, the re-nudge after it, the four-hour boundary, per-home independence, and that the ask stays out of a re-ring ladder that still rings an ordinary steer beside it. The obsolete observation-ordering test is deleted with the machinery it covered. * no-mistakes(review): Serialize reconcile cooldown commits with mate lifecycle * no-mistakes(review): Reject stale reconcile snapshots across mate reincarnations * no-mistakes(review): Start reconcile cooldown after delivery completes * no-mistakes(review): Keep reconcile sends nonblocking and remove pending residue * no-mistakes(document): Document reconcile skip and stale-endpoint behavior * no-mistakes(lint): Fix reconcile test subshell lint warning * no-mistakes: apply CI fixes * fix(bin): reconcile markerless remote secondmates safely (#3140) * fix(bin): stop dropping reconcile nudges for markerless remote secondmates A persistent remote secondmate's parent-side state/<id>.meta never carries spawn_gen: bin/fm-spawn.sh's spawn_remote_secondmate() is its sole writer and never writes one, because that incarnation identity does not apply to a remote route. fm-secondmate-reconcile.sh's row filter required a non-empty spawn_gen matching an identifier regex, so every such row was silently dropped before the per-row loop ever saw it: no sent/stale/failed line, no cooldown record, nothing sent, and no trace of why. Give a legitimately markerless persistent remote secondmate a safe substitute identity - its recorded remote_host - instead of weakening the spawn_gen check for rows that do have a generation: - bin/fm-secondmate-reconcile.sh: carry host through the row projection for both fm-fleet-snapshot.v1 and fm-bearings.v1 documents, and admit an empty spawn_gen instead of filtering the row out. A new revalidate_identity() compares the sampled spawn_gen against current metadata when one was sampled (unchanged), or the sampled host against the metadata's remote_host when none was sampled and the metadata still carries no spawn_gen of its own. A row with neither a spawn_gen nor a host has no safe identity at all and fails loudly instead of vanishing, exactly the visibility the original bug lacked. - Rows now join on the ASCII unit separator rather than @tsv: bash's IFS-whitespace read collapses consecutive tabs, which would have silently dropped a legitimately empty field again. - bin/fm-bearings-snapshot.sh: thread host through the secondmate_reconcile projection so the fm-bearings.v1 path (the one bearings itself feeds to the reconcile hook) carries the same substitute identity. - tests/fm-secondmate-reconcile.test.sh: end-to-end coverage through the real remote transport (fm-on.sh + fm-remote-secondmate-control.sh against a genuinely seeded remote home) for a markerless mate nudged once per cooldown window, a stale/replaced remote route refused exactly like the existing local spawn_gen case, and a row with no identity at all failing loudly rather than being swallowed. * no-mistakes(review): Enforce markerless remote host identity during final delivery * no-mistakes(document): Document markerless remote reconciliation safety * fix(bin): hand a busy declared pause to the away-mode daemon once, undecorated (#3147) * fix(watch): hand a busy declared pause to the away-mode daemon undecorated While away mode is active the daemon owns triage and the watcher reverts to one-shot, handing over plain wake identities the daemon classifies itself. The busy-turn bound was the one stale path that did not: with afk active it ran the wedge timer, so the daemon received a wake already decorated as a possible wedge. That decoration outranks the daemon's own verdict. handle_wake escalates an enriched wedge reason before its pause classification can apply, so a crew that declared the wait itself - a `paused:` external wait or a verified captain-held transfer holding a live foreground call - was wedge-escalated once per FM_STALE…
geojitsu
added a commit
to geojitsu/firstmate
that referenced
this pull request
Sep 4, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-queue fallbac…
withally
added a commit
to withally/firstmate
that referenced
this pull request
Sep 4, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-queue fallback …
lytv
pushed a commit
to lytv/mymate
that referenced
this pull request
Sep 8, 2026
…ath (kunchenguid#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes
doitdigital0495
added a commit
to doitdigital0495/firstmate
that referenced
this pull request
Sep 9, 2026
…wake rows (#14) * fix(bearings): restore decision options and add close controls (#2707) * fix(bearings): always show decision options and a close/drop control Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call. * no-mistakes(review): Fix drop closure and decision-only option validation * no-mistakes(review): Preserve answerability for non-decision cards * no-mistakes(document): Clarify decision drop documentation * ci: require no-mistakes pipeline step attestation (#2710) Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed. * feat: collapse decisions into tasks held for the captain (#2728) * feat(captain-hold): collapse the decisions concept into tasks held for the captain A decision is no longer a separate type: it is an ordinary backlog task held for the captain, identified by its task id. bin/fm-captain-hold.sh owns the surviving behaviors - guarded hold creation, the recorded-answer close (answer/answers with a release mode for captain-gated work), the source bindings, and the investigation completion gate - and bin/fm-decision-hold.sh becomes a one-release compatibility shim over it. The fleet snapshot now parses hold-until and computes captain_actionable as queued + captain-held + unblocked + due, independent of row kind, plus a presentation-only deferred_marker for prose-deferred rows. Bearings renders every due captain-held task in Captain's Call, date-deferred holds as dated Charted Next gates, suppresses prose-deferred rows from default views with an omitted disclosure, and excludes from Recently Landed anything that closed while still held for the captain. Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already plain task ids and keep working; short keys in recorded metadata, concrete origin bindings, chat --resolve-key fallbacks, and old resolution records all resolve in place. * no-mistakes(review): Fix captain answer replay and body preservation * no-mistakes(review): Fix captain hold idempotency and legacy replay * no-mistakes(review): Validate card close modes and compatibility routing * no-mistakes(review): Enforce release replay mode matching * no-mistakes(review): Prevent duplicate decision cards and released replay mismatches * no-mistakes(review): Preserve answer columns and legacy resolve replays * no-mistakes(document): Document strict replay and legacy compatibility * no-mistakes(lint): Quote done literals to satisfy ShellCheck * no-mistakes: apply CI fixes * fix(rebase): keep collapsed captain hold board semantics * fix: bound recovery announcements and preserve supervision (#2733) * fix(watch): announce recovery once per generation and keep successors supervising A lost Pi/OpenCode handling handshake re-announced the same recovery generation on every cycle and spent the successor's first ~55s blind, so a real crew event could be ignored and then dropped. Record the announcement in the durable marker, confirm the handshake before the follow-up without swallowing failure, and enter the poll loop immediately. * no-mistakes(review): Tighten recovery event timing regression * no-mistakes(document): Document recovery-loop supervision guarantees * fix(bin): surface captain-call record divergence (#2744) * fix(bin): signal a captain call resolved in the log but still held A captain call has two records and closing one has never closed the other: a `resolved [key=...]` line closes the status-log fold, while the backlog task held for the captain closes only through `fm-captain-hold.sh answer`. Answering on the status side alone left no trace of the disagreement - the fold went quiet, the durable record kept saying the captain owed an answer, and nothing warned. The defect was never the separation; it was the silence. Add `fm-captain-hold.sh diverged`, a read-only report of that contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one condition: a task still open and still carrying the captain-hold annotations whose key was closed on the status side by the resolve verb, under the collapsed identity or the legacy derived one. It closes nothing, ever. A captain call closed wrongly leaves review entirely, which is worse than the noise, so both reconciliation directions stay human-owned and the printed hint names both - a resolution is not proof the captain ruled, since a call can dissolve on a false premise or turn out to have been a question of fact. Three states are deliberately not divergence: a `captain-held` close is the verified transfer `complete` writes, a still-open keyed decision belongs to the OPEN DECISIONS fold, and a captain call with no routed work item is legitimate rather than incomplete, so routed work is no part of the test. `fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how the status side currently reads one key by replaying the existing `_fm_decision_fold_line` rule rather than re-deriving it, so the two closing verbs stay distinguishable in one place. The per-wake cost is one `tasks-axi list`, one key scan per status log, and the precise per-key fold only for a key that already names a still-open task; the call is hard-bounded so a slow backlog tool can never delay wake presentation. * fix(document): Correct divergence lifecycle documentation * fix(document): Neutralize divergence lifecycle prose * fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524) * fix(watch): re-arm supervision after an abandoned auto-arm claim A Claude auto-arm cycle that armed, delivered one rewake, and exited left its single-flight lock behind. Both Stop-event participants then deferred to that lock forever, because its recorded pid was still live: the turn-end guard read it as recovery under way and allowed the stop, and the next Stop firing treated it as another owner and declined to arm. On 2026-08-14 a home with two tasks in flight lost supervision for about 40 minutes with no watcher process and no watcher lock, its beacon frozen at the one delivery, and both crewmates' finished reports sat in the durable queue until an operator drained it by hand. Abandonment is now proven from the epoch ledger instead of inferred from pid liveness. A lock whose holder pid matches the ledger's own owner_pid while the recorded outcome is anything other than arming has already finished its decision, so that claim is reclaimed under the lock's steal mutex, stops counting as recovery ownership in the guard, and is cleared by the guard's terminal check rather than deferred to. A failed clear re-blocks instead of allowing a blind stop, and an arming entry stays in flight however old it is, because its owner foregrounds the arm for the whole watcher cycle. Issue #2251's PR #2263 does not cover this failure. It is closed and unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled watcher and matching stale watcher lock of an arm that is currently running. Here no arm and no watcher were running and no watcher lock existed, so it has nothing to retire and the home stays blind. tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming and unnamed-owner cases that must keep the gate closed, and the failed clear. tests/fm-turnend-guard.test.sh covers the guard side of the same boundary. Both fail without this change. * fix(watch): defer a wedge escalation while the task worktree is written The wedge detector had two inputs, rendered pane quietness and the run step, and neither can see a crew that is writing source, then tests, then documentation behind a static pane. On 2026-08-14 one crewmate produced eight consecutive possible-wedge escalations in a single afternoon, three of them demanding deep inspection, while it was demonstrably working and then committed. Every one of them cost a supervision turn to disprove by hand. Add write activity inside the crew's own recorded worktree as a third liveness input. crew_worktree_written_since compares the worktree against the caller's existing idle-window timer file, so -newer needs no clock arithmetic, no temp file, and no portable mtime write. The probe runs only inside the branch that was about to escalate, which bounds it to one pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and leaves the per-poll stale sweep exactly as cheap as before. Positive evidence defers rather than cancels. The idle timer restarts so the next window probes again, the escalation counter is neither advanced nor reset so a later genuine wedge keeps the demand-deep-inspection history it earned, and a .writing-since marker ages the whole deferral chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS, through the same throttle shape a declared pause already uses, labeled as a recheck rather than a wedge. This can only reduce false positives: every absence of evidence, including no recorded worktree, a torn-down worktree, a missing anchor, and a failed walk, falls through to the unchanged escalation schedule, so a crew that writes nothing still escalates on the existing timetable. What the signal cannot see, by design or by construction: - CPU burn with no writes, such as a long compaction, is invisible. That case keeps the old behavior exactly. - A commit-only phase writes only .git, which is pruned first so that firstmate's own read-only git commands against the worktree can never make the probe self-fulfilling. - Writes under the pruned generated trees, or deeper than FM_WORKTREE_WRITE_MAXDEPTH, do not count. - The probe cannot attribute a write to the crew, so a background build or another process touching the tree looks the same. The hourly re-surface is what bounds that, and a churny file cannot buy silence. - The away-mode daemon's own escalation path is deliberately untouched. tests/fm-watch-triage.test.sh covers the classifier including the .git prune, both halves of the live case on one fixture (quiet plus writing defers, quiet plus silent still escalates and counts), and the bounded re-surface. All three fail without this change. * no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe * no-mistakes(document): document away-mode wedge boundary and probe filesystem limit * no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims * fix(watch): keep a write deferral scoped to its own idle window Two consistency gaps in the worktree write probe, both found while reviewing the wedge-deferral change on this branch. A write deferral is a bounded chain: its .writing-since marker ages the whole chain so a churning worktree still re-surfaces once per resurface window. That is only sound while the chain belongs to the current quiet stretch, so every path that restarts the idle-window timer has to drop it too. Two did not: the corrupt-timer repair in wedge_timer_check, and both first-sight branches for a captain-relevant status. A chain left over from an earlier quiet stretch made the first deferral of the new window re-surface immediately instead of after a full fresh window. FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing" and is the obvious way to widen the probe to the whole depth-bounded tree. Instead an empty list reported no evidence at all, quietly costing the wedge detector its third liveness input on a home that meant to widen the walk. An empty list now widens the walk, and the header says so. Neither change alters when a stall that writes nothing escalates. Regressions in tests/fm-watch-triage.test.sh cover all three paths and each one fails on the pre-fix code. * no-mistakes(review): honor an empty write-prune, bound the probe, share window_key * no-mistakes(document): align probe knob count and guard regression-coverage ownership * no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test * fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748) * fix(bin): give a captain hold the same bounded pause cadence as a declared pause Two supervisors read a finished task's last status line and disagreed about which declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh suppresses its inactive-outcome scan only on `captain-held`, while the away-mode daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so the two verbs are mutually exclusive and no finished task waiting on a person could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the 900s outcome scan and immediately produced five possible-wedge escalations in one batch, because the 240s wedge detector no longer saw a pause verb. fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This extends that same answer to the paths still asking the narrower one: - bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to move together. classify_stale returns the pause action, reconcile_pause_tracking and migrate_watcher_pause_markers record and migrate the marker, and housekeeping defers the wedge and then re-surfaces the recheck. Changing only the stale-persistence gate would defer the escalation while reconcile_pause_tracking recorded nothing, so the wedge marker would persist and the sweep would `continue` past it forever: quiet, but never re-surfacing. - bin/fm-watch.sh's secondmate stale gate, whose downstream owner pause_state_class already treats both declarations identically. - bin/fm-push-transition-lib.sh's absorb, where either declaration already names the human the transition would report and the wait is already durably recorded. Quieting alone would be half a fix, so the bounded re-surface had to reach a hold too. A hold has no current-state mapping, unlike `paused`, so authoritative crew state reports it as unknown and pause_state_class received `none`. An ordinary crew recovers pause classification from that state through confirmed agent death, which proves no live decision gate is being silenced. A secondmate's endpoint liveness is deliberately never read there, because an idle mate is healthy by design, so that confirmation is unavailable by construction and cannot be required: without recovering the classification for a mate, every caller silenced a held mate outright and its hold would rot invisibly. That promotion is bounded by the declared-wait guard at the top of the function, so it can only reclassify a task that already declared a wait and shows no positive working evidence. Two narrow `status_is_paused` calls are deliberately left alone. bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a wedge path; reporting a hold as `paused` would erase the distinction status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held` close is a verified durable transfer and a `resolved` close claims outright settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no change because that function's own case list already returns non-relevant for `captain-held`. bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its guard exists because a finished task's crew state still reports done from a higher-priority source than the log, and a declared pause needs no such guard: the scan only reports done or failed, and nothing else reaches its record path. Widening it would change a separate subsystem's reporting contract, which this defect does not require. Coverage extends the existing colocated patterns for these predicates and asserts both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker converting to pause tracking with no escalation, the bounded re-surface with its window reset, and the boundary case where an answered hold stops claiming the cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on the same bounded cadence without being labeled a wedge. tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one of these fails on the pre-fix code except the answered-hold boundary case, which is there to pin that the quieting was not widened too far. The `paused:` workaround appended to those 11 tasks is live supervision state and is untouched here. It can be retired once this lands. * no-mistakes(review): name the captain in a held task's bounded recheck * no-mistakes(document): extend declared-wait supervision docs to captain-held holds * fix(bin): make lint prerequisites and harness tests reliable (#2758) * fix(lint): name the installer when ShellCheck or actionlint is missing A missing actionlint exited 127 like a bare command-not-found. Fail with exit 1 and point at the pinned installer, matching the missing-ShellCheck path, without weakening the version pin. * test: isolate kimi and muse detection from inherited Cursor markers Harness detection checks CURSOR_AGENT before ancestry, so these markerless-adapter cases failed when the suite itself ran under Cursor. Clear the verified markers the same way the secondmate harness tests already do. * no-mistakes(document): Document Muse Cursor marker cleanup * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after #2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (#2857) * feat(bin): deliver local steers through durable task inboxes (#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes #2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (#2939) * fix(bin): bound remote job worker supervisor restarts (#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix: safely split supervision wake handling by actor (#2953) * feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup Three related fixes to the shared wake-drain and Pi supervision-branch dispatch machinery so a routine success is never main-blocking and a mixed queue can safely split between actors. 1. Successful routine results no longer create main-blocking wake rows. fm-startup-network.sh only enqueues a check: startup-network wake when the deferred result is actionable (state is not "done", or the report carries a bootstrap-diagnostics actionable prefix); a clean success stays durable in the report file without ever waking the agent. 2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes presentation and --ack-through to the current actor (bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original whole-queue cutoff behavior, unaffected. A branch actor (FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision branch's own bash tool calls) is scoped to an explicit eligible-row snapshot instead of a cutoff comparison, so it can never remove a row it was not granted - the fix for the swallow risk that used to force an all…
doitdigital0495
added a commit
to doitdigital0495/firstmate
that referenced
this pull request
Sep 11, 2026
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task inboxes
Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote
secondmate steer now crosses fm-on.sh as a durable record written
idempotently into the remote home's steering inbox plus a best-effort
remote doorbell, and the last typed-payload steer transport is deleted:
- fm-remote-secondmate-control.sh cmd_send writes the record via the new
fm_task_inbox_write_idempotent and rings the doorbell; it no longer
types the payload through an inner fm-send at an explicit pane target.
- fm-send.sh routes every remote text steer (harness-native included,
which marking already reduced to chat) onto the remote inbox leg,
retries the identical leg once on ssh 255, closes --resolve-key
decisions at enqueue for remote too, and preserves a marked request's
reply expectation when completion stays unknown. The exit-3-as-
delivered remap, the 255 do-not-resend trap, and the remote typed
submit block are removed.
- fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run
lands on the existing record, handled or not, so an ambiguous
transport can always be safely re-run.
- Tests pin the new contract end to end (record + doorbell + no typed
payload across ssh, one-record idempotence under an ambiguous
transport, enqueue-time decision close, loud real failures, and the
deleted typed-payload behaviors gone), and AGENTS.md plus
docs/remote-secondmates.md describe the remote leg's new semantics.
* no-mistakes(review): Harden remote inbox delivery against lifecycle races
* no-mistakes(review): Enable correlation-preserving remote steer resends
* no-mistakes(review): Fail closed on stale correlation resends
* no-mistakes(review): Include home context in remote resend commands
* no-mistakes(review): Lock and revalidate remote parent routes
* no-mistakes(document): Clarify remote steer retry documentation
* no-mistakes: apply CI fixes
* feat: add persistent Pi supervision branch (#2858)
* wip: forked supervision on Pi (checkpoint before docs)
* fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement
Peek-then-shift mirror flush so a failed append retries instead of dropping;
durable mirror cursor commits only after delivery into the branch;
the main fallback wake is operational-encoded like every watcher injection;
session_shutdown quiesces the generation and session_start re-arms, so /new
and /resume no longer kill the branch permanently. Registers the extension in
the strict typecheck, adds the dispatch handshake test, the branch extension
suite, the bash-level regression suite, the session-start replay test, and
the opt-in real-SDK live guard.
* test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures
The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown
sources fm-lease-lib.sh, so every fixture that copies or symlinks those
files in isolation gains the new sibling.
* no-mistakes(review): Prevent shutdown wake loss and serialize lease claims
* no-mistakes(review): Durably hand off wakes and retain portable leases
* no-mistakes(review): Require durable reports and clear disposed branch leases
* no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup
* no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries
* no-mistakes(review): Require complete acknowledgements and replay cleanup failures
* no-mistakes(review): Bind supervision to lock ownership and durable delivery
* no-mistakes(review): Activate branch lazily after session lock acquisition
* no-mistakes(review): Preserve undelivered mirror context across extension rebinds
* no-mistakes(review): Acknowledge startup replay only after main delivery
* no-mistakes(review): Isolate replay metadata from untrusted digest content
* no-mistakes(review): Reject duplicate reports for active wake sequences
* no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay
* no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts
* no-mistakes(review): Anchor wake sequence matching to outcome fields
* no-mistakes(document): Clarify Pi supervision durability contracts
* no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* refactor(pi-branch): collapse to confused-agent-grade guards per captain decision
Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat
model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade
separation is impossible in the shared-process design and is filed as
separate follow-up work. Rip out the machinery that chased it: the
generation fence and shell-provenance markers, the wrapper-tagged ancestry
walks, guard auto-claim with per-script release traps, the pending-wake
files and ack-receipt correlation (the durable wake queue already
re-presents anything unacknowledged), the delivery-receipt store with
contiguous cursor advancement, the session-start replay-metadata channel,
and the branch tool quiescence counters.
Keep the behaviors the board requires, each on its simplest implementation:
lazy per-action session-lock ownership (cold start activates after the lock
lands; a secondary session stays inert), mirror durability across extension
rebinds via the durable cursor, replay-exactly-once from the one read
cursor, the awaited operational-encoded fallback, per-generation stray-lease
cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home
never honors a leftover lease), the loud accidental-override guards
(readonly actor prelude, cross-actor claim refusal), and the role-partition
refinements (no forced teardown, no direct relaunch for the branch).
Default-on-for-Pi is unchanged.
* no-mistakes(review): Enforce lock ownership and serialize lease mutations
* no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner
* no-mistakes(review): Report outcomes before acknowledging durable wakes
* no-mistakes(review): Restrict leases to Pi and instruct main claims
* no-mistakes(review): Reject malformed lease locks and torn outcome tails
* no-mistakes(review): Validate complete outcome tails before appending
* no-mistakes(review): Guard branch side effects across session replacements
* no-mistakes(document): Update Pi supervision durability and lease documentation
* no-mistakes(lint): Suppress intentional nested-shell expansion warning
* no-mistakes: apply CI fixes
* fix(pi-branch): authorize lease releases by caller
* fix(lint): break redundant source-analysis path in fm-lease-lib.sh
fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's
--external-sources traversal a second path into an already 1540-line file
that fm-send.sh and fm-teardown.sh also source directly, blowing up the
recursive analysis past CI's lint timeout. Mark it a source=/dev/null
analysis boundary, matching the existing fm-task-inbox-lib.sh convention.
Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared
serial-lint definition (dropping an unrelated parallel-sharding change
that was itself hanging and masked this root cause).
* no-mistakes(document): Correct lease caller-authorization documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): parallelize startup network sweeps (#2927)
* feat(bin): parallelize session-start remote secondmate network sweeps
Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync.
Co-authored-by: Cursor <cursoragent@cursor.com>
* no-mistakes(document): Document parallel startup network sweeps
* no-mistakes(lint): Fix empty environment assignment lint warning
* no-mistakes: apply CI fixes
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* test: handle absent watcher wake queues (#2845)
* fix(tests): count declared-pause wakes without crashing on an absent queue
The exited-declared-pause case counts queued stale wakes by handing
state/.wake-queue straight to awk. A watcher that queues nothing never
creates that file, and awk aborts on a missing path before its END rule
runs, so the count collapses to the empty string. The next comparison
then fails as an integer-expression error and surfaces as a wake flood
with no number, hiding the real contract breach the following grep names.
Read the queue the way the drain-count assertion at the end of this file
already does: silence awk's open error and default an absent queue to
zero. Applied to all four counts in this case, including the live
external-decision gate pair whose queue an acknowledged drain can also
leave behind. An absent queue now reports "did not use the bounded
paused recheck", while a genuine flood still fails with its real count.
Fixes #2628
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* style(pi): distinguish routine and captain supervision merge notes by icon (#2934)
* style(pi): restyle supervision merge notes with a sailboat and matching pad
Secondary-session notes were flush against the TUI edge and fully tinted.
Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages.
* style(pi): distinguish routine and captain merge notes by icon only
Visible notes now lead with a sailboat or anchor, then only the dim outcome.
Drop the branch-merged wording and verdict brackets so the icon is the only kind signal.
* docs(pi): add the approved multi-brain architecture poster (#2938)
The markdown contract stays the owner; the still is only the visual of the idea.
* feat(pi): default branch supervision and route heartbeats (#2939)
* fix(bin): bound remote job worker supervisor restarts (#2942)
* fix(bin): bound remote worker supervisors
* no-mistakes(review): release incumbent supervisor before starting its replacement
* no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it
* no-mistakes(review): narrow remote worker change to restart accounting only
* no-mistakes(document): clarify supervisor restart guard is a lifetime total
* fix: safely split supervision wake handling by actor (#2953)
* feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup
Three related fixes to the shared wake-drain and Pi supervision-branch
dispatch machinery so a routine success is never main-blocking and a
mixed queue can safely split between actors.
1. Successful routine results no longer create main-blocking wake rows.
fm-startup-network.sh only enqueues a check: startup-network wake when
the deferred result is actionable (state is not "done", or the report
carries a bootstrap-diagnostics actionable prefix); a clean success
stays durable in the report file without ever waking the agent.
2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes
presentation and --ack-through to the current actor
(bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original
whole-queue cutoff behavior, unaffected. A branch actor
(FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision
branch's own bash tool calls) is scoped to an explicit eligible-row
snapshot instead of a cutoff comparison, so it can never remove a row
it was not granted - the fix for the swallow risk that used to force
an all-or-nothing whole-…
LorenzoMinghini
added a commit
to LorenzoMinghini/firstmate
that referenced
this pull request
Sep 15, 2026
* test(procevent): make the process-event suite's detached-runner assertions deterministic (#2617)
Three assertions in tests/fm-procevent.test.sh depended on a detached runner
having finished work that the command starting it does not wait for.
reconcile's replacement runner is started through detach_runner, which only
forks: reconcile returns and counts the start before that runner has claimed
its source or exec'd its child. Any assertion taken straight after reconcile
therefore samples a race.
- The publish-before-apply recovery section left its always-ready /bin/echo
source registered across the recovery reconcile, so that reconcile launched
a competing detached poll (observed: started=1) that then raced every later
assertion for the source claim, the next capture sequence, and this home's
applied record, and outlived the section holding a live claim. It is now
retired before that reconcile - re-announcement is proven from the durable
inbox alone and needs no registration - and started=0 is asserted so a
competing poll cannot be reintroduced unnoticed. This is the same
retire-before-reconcile discipline the self-announcing section already
carries; that section acquired it after the identical race made its
"not-autohandled: self-src" assertion read "already owned: self-src".
- The crashed-leader replacement section snapshotted the replacement's claim
file and execution log behind a fixed 0.5s settle window. On a loaded
machine that window expires first, which is the CI flake behind "a
replacement runner started without recording its own claim" and "reconcile
did not start exactly one replacement source". Both effects are now waited
for with the suite's bounded wait helpers; the exact one-replacement count
is still asserted afterwards, unchanged.
- The duplicate-start section slept 0.5s for reconcile's runner to record
ownership before asserting that a second start loses to it. It now waits
for that claim.
Also tighten one assertion that could not fail as written: "autohandled:
self-src" is a substring of "not-autohandled: self-src", so the applied path
was accepted even when the runner reported the capture left for the handler.
Evidence: on the unmodified suite, 128 full runs at 6-8x concurrency produced
6 failing runs, all in the crashed-leader section. On the fixed suite, 216
full runs under the same load produced none. Reverting the self-announcing
section's retire-before-reconcile line reproduces "already owned: self-src"
on the first iteration, confirming the shared mechanism.
* fix: preserve pending replies and defer remote reposts (#2618)
* fix(bin): keep pending-reply expectations honest on both send legs
Two related asymmetries let the parent-owned secondmate reply guard drop or
nag requests it should not have.
Local delivered-unconfirmed dropped the expectation. A marked request whose
submit read-back stayed unconfirmed (verdict=pending) is the same
not-a-failure outcome the remote leg reports as delivered, but fm-send
discarded the parent's pending-reply record for it, so a request that very
likely landed stopped being tracked entirely. The record now stays armed on
its unconfirmed-delivery marker: a correlated report still resolves it, and
an unanswered one still surfaces through the library's own reconciliation.
Exit 3 and the local rule that an unconfirmed answer never closes a decision
key are unchanged.
Remote replies were nagged for a repost they did not need. A remote mate's
report reaches the parent's status log only through the asynchronous mirror
in fm-procevent-remote-reply.sh, yet the guard read an absent correlated
line as proof the mate never reported - even while the answer was still in
flight, which is the common case because the mirror's poll window is
comparable to the recovery grace. The mirror now publishes one caught-up
watermark from a quiet window, and the guard admits a missing report as
evidence only once that watermark passes the turn that should have produced
it. A genuinely missed report still gets exactly one repost, and a channel
that is behind, unarmed, or broken leaves the request durably open and
un-nagged rather than nagging blind; the mirror escalates its own continuity
failures as before.
Tests: a local unconfirmed secondmate send keeps its expectation armed and
resolvable; a mirrored correlated remote reply resolves with no repost; a
stale or absent watermark withholds the repost while a fresh one still
releases it; a quiet remote window publishes the watermark and retirement
clears it.
* no-mistakes(review): Distinguish preempted polls from quiet windows
* no-mistakes(document): Clarify remote reply channel freshness
* no-mistakes(lint): Annotate shared remote preemption exit constant
* fix(bin): honor declared pauses in busy-pane wedge checks (#2619)
* fix(watch): honor a declared pause on a busy pane's completed-turn bound
A worker that declares an external wait (`paused:`) and then blocks in one
long foreground call - a review-hosting scout parked in a single blocking
`lavish-axi poll`, a bounded watch loop, a rate-limit sleep - keeps its pane
BUSY, so the stale path that already honors declared pauses never ran for it.
The busy-pane completed-turn bound instead routed it straight into
wedge_timer_check, which re-escalated "possible wedge, escalation N" (and, past
the threshold, demand-deep-inspection) every FM_STALE_ESCALATE_SECS for as long
as the review stayed open.
busy_turn_bound_check now owns which absorber takes a crossed bound: a crew
whose own last status line declares an external wait or a verified captain-held
transfer takes the bounded FM_PAUSE_RESURFACE_SECS recheck, and everything else
keeps the unchanged wedge timer. The discriminator is the declaration together
with liveness (the caller has already confirmed the pane is busy), never a
blanket silencing - a crew that declared nothing, or whose pane is not live,
escalates exactly as before, and a declared pause still re-surfaces once per
long cadence so a forgotten wait cannot rot invisibly. Away mode is untouched:
the daemon owns pause triage there and already reads the same vocabulary.
The two call sites also no longer clear pause bookkeeping in the same poll the
pause cadence recorded it, which would have erased the re-surface throttle and
turned the long cadence back into a per-poll re-surface.
Tests: a three-phase regression fixture pins the absorbed pause, its long-cadence
recheck, and the restored wedge escalation once the declaration is lifted on the
same busy over-age pane.
Also de-flakes tests/fm-watch-triage.test.sh, which failed spuriously on a loaded
machine: fixed liveness budgets were reaping watchers mid-startup, so assertions
on post-poll state passed vacuously or failed spuriously. Waits that describe a
poll's outcome now wait for a completed poll cycle via the liveness beacon, the
heartbeat test waits for the heartbeat it asserts on, and every wait_for_exit
budget is the uniform 10s already used elsewhere in the file.
* no-mistakes(review): Fail poll-cycle waits on timeout
* no-mistakes(review): Prevent poll timeout test hangs
* no-mistakes(document): Clarify paused busy-pane supervision
* doc: Enhance communication guidelines for decision-making
Added guidelines for decision communication to the captain.
* doc: Update task delegation and communication guidelines
Clarify communication protocols with crewmates regarding task delegation and reporting.
* fix(bin): reliably confirm herdr steer submission (#2647)
* fix(herdr): confirm local steers that native agent-state misses
Herdr can leave agent_status idle for a landed Claude turn and can keep
queued Enter text visible while busy, so fm-send was reporting false
swallows. Confirm those cases through the shared queued-Enter verdict
and a cleared composer, and keep a genuine idle pending composer as
unconfirmed.
* no-mistakes(review): Stop Herdr Enter retries on unreadable composers
* no-mistakes(review): Reject queued delivery when all Herdr Enter sends fail
* no-mistakes(review): Prevent confirmation after failed Herdr Enter
* no-mistakes(review): Pace Herdr retries and clarify submit fallback
* no-mistakes(review): Align Herdr submit docs with idle fallback
* no-mistakes(document): Correct Herdr submit-confirmation documentation
* feat(bearings): add interactive Lavish fleet board (#2659)
* feat(bin): accept any-origin decision bindings with full-identity keys
An aggregation surface (the bearings board) carries captain answers for holds
across origins, but a binding was one-origin-per-source and the Lavish adapter
capped question keys at 64 chars while real full hold identities measure 69-81.
- fm-decision-hold.sh: bind <source-id> --any-origin records the (any) marker;
binding prints it verbatim and answers accepts it, so the runner's feed seam
carries an any-origin source with no runner change. In any-origin mode each
key is a full hold identity <origin>-decision-<key>, split at its first
-decision-; a key with no separator (merge/dispatch instructions) is skipped
and feeds nothing, keeping non-decision answers out of the hold ledger by
construction. Every existing close guard applies unchanged.
- fm-procevent-lavish.sh: raise the question-key cap 64 -> 128 so a full hold
identity fits; the slug-shape security property is unchanged.
- tests: cross-origin closure through the real runner seam, an 81-char
identity through the adapter, cap and shape refusals, routed-work skips,
nonexistent-identity skips, and idempotent replay.
* feat(bearings): add the /bearings lavish interactive fleet board
/bearings lavish renders the bearings snapshot onto a shipped, reusable board
template and arms it as a Lavish process-event source, so the captain answers
Captain's Call items on the board and firstmate is woken by an ordinary check
wake - no conversational turn ever blocks on a poll.
- .agents/skills/bearings/assets/board-template.html: the shipped template
(myfirstmate design system inlined, one fm-bearings-board.v1 JSON slot,
fail-closed schema guard that renders an error card instead of an empty
fleet). Per-invocation agent work is composing the payload only.
- bin/fm-bearings-board.sh: build/refresh owner - fail-closed payload
validation, slot injection with a round-trip check and \u003c escaping,
stable board path, any-origin bind ALWAYS before arm, arm-if-absent.
- bearings SKILL.md: the lavish invocation option, board composition rules,
board-wake handling, and the captain-ruled merge-click authorization with
its mandatory safeguards (PR resolved from the task's own meta record,
wake-time green re-verification, never a red or changed PR, merges only
through bin/fm-pr-merge.sh, chat echo with the full PR URL).
- process-event-sources SKILL.md: one-line board-wake routing trigger.
- tests: payload refusals, injection round-trip, bind-before-arm, idempotent
re-arm, and template slot integrity.
Fleet pickup: homes receive this after merge plus a firstmate self-update;
landing timing is coordinated with the main firstmate.
* no-mistakes(review): Harden bearings board validation and wake handling
* no-mistakes(review): Require HTTPS for bearings board PR links
* no-mistakes(review): Fail closed and bound bearings board answers
* no-mistakes(review): Enforce UTF-8 byte limits for board answers
* no-mistakes(review): Serve bearings board before arming and reject empty actions
* no-mistakes(review): Prove bind-before-arm ordering through live answer consumption
* no-mistakes(document): Document bearings board and cross-origin answers
* fix(bearings): restore decision options and add close controls (#2707)
* fix(bearings): always show decision options and a close/drop control
Freeform-only Captain's Call cards hid the option buttons the board was designed around, and there was no way to drop a stale hold without inventing an answer. Require selectable options, keep freeform as a supplement, and route the reserved __drop__ answer through decline so the hold leaves Captain's Call.
* no-mistakes(review): Fix drop closure and decision-only option validation
* no-mistakes(review): Preserve answerability for non-decision cards
* no-mistakes(document): Clarify decision drop documentation
* ci: require no-mistakes pipeline step attestation (#2710)
Signature-only PRs can hide skipped review, test, or document steps. Fail unless no-mistakes >= 1.46.0 attests those three steps completed.
* feat: collapse decisions into tasks held for the captain (#2728)
* feat(captain-hold): collapse the decisions concept into tasks held for the captain
A decision is no longer a separate type: it is an ordinary backlog task held
for the captain, identified by its task id. bin/fm-captain-hold.sh owns the
surviving behaviors - guarded hold creation, the recorded-answer close
(answer/answers with a release mode for captain-gated work), the source
bindings, and the investigation completion gate - and bin/fm-decision-hold.sh
becomes a one-release compatibility shim over it.
The fleet snapshot now parses hold-until and computes captain_actionable as
queued + captain-held + unblocked + due, independent of row kind, plus a
presentation-only deferred_marker for prose-deferred rows. Bearings renders
every due captain-held task in Captain's Call, date-deferred holds as dated
Charted Next gates, suppresses prose-deferred rows from default views with an
omitted disclosure, and excludes from Recently Landed anything that closed
while still held for the captain.
Legacy compatibility: pre-collapse <origin>-decision-<key> rows are already
plain task ids and keep working; short keys in recorded metadata, concrete
origin bindings, chat --resolve-key fallbacks, and old resolution records all
resolve in place.
* no-mistakes(review): Fix captain answer replay and body preservation
* no-mistakes(review): Fix captain hold idempotency and legacy replay
* no-mistakes(review): Validate card close modes and compatibility routing
* no-mistakes(review): Enforce release replay mode matching
* no-mistakes(review): Prevent duplicate decision cards and released replay mismatches
* no-mistakes(review): Preserve answer columns and legacy resolve replays
* no-mistakes(document): Document strict replay and legacy compatibility
* no-mistakes(lint): Quote done literals to satisfy ShellCheck
* no-mistakes: apply CI fixes
* fix(rebase): keep collapsed captain hold board semantics
* fix: bound recovery announcements and preserve supervision (#2733)
* fix(watch): announce recovery once per generation and keep successors supervising
A lost Pi/OpenCode handling handshake re-announced the same recovery
generation on every cycle and spent the successor's first ~55s blind, so
a real crew event could be ignored and then dropped. Record the
announcement in the durable marker, confirm the handshake before the
follow-up without swallowing failure, and enter the poll loop immediately.
* no-mistakes(review): Tighten recovery event timing regression
* no-mistakes(document): Document recovery-loop supervision guarantees
* fix(bin): surface captain-call record divergence (#2744)
* fix(bin): signal a captain call resolved in the log but still held
A captain call has two records and closing one has never closed the
other: a `resolved [key=...]` line closes the status-log fold, while the
backlog task held for the captain closes only through
`fm-captain-hold.sh answer`. Answering on the status side alone left no
trace of the disagreement - the fold went quiet, the durable record kept
saying the captain owed an answer, and nothing warned. The defect was
never the separation; it was the silence.
Add `fm-captain-hold.sh diverged`, a read-only report of that
contradiction, and print it from `fm-wake-drain.sh` as a bounded RECORD
DIVERGENCE section beside OPEN DECISIONS on every drain. It flags one
condition: a task still open and still carrying the captain-hold
annotations whose key was closed on the status side by the resolve verb,
under the collapsed identity or the legacy derived one.
It closes nothing, ever. A captain call closed wrongly leaves review
entirely, which is worse than the noise, so both reconciliation
directions stay human-owned and the printed hint names both - a
resolution is not proof the captain ruled, since a call can dissolve on a
false premise or turn out to have been a question of fact.
Three states are deliberately not divergence: a `captain-held` close is
the verified transfer `complete` writes, a still-open keyed decision
belongs to the OPEN DECISIONS fold, and a captain call with no routed
work item is legitimate rather than incomplete, so routed work is no part
of the test.
`fm-classify-lib.sh` gains `status_key_closing_verb`, which reports how
the status side currently reads one key by replaying the existing
`_fm_decision_fold_line` rule rather than re-deriving it, so the two
closing verbs stay distinguishable in one place. The per-wake cost is one
`tasks-axi list`, one key scan per status log, and the precise per-key
fold only for a key that already names a still-open task; the call is
hard-bounded so a slow backlog tool can never delay wake presentation.
* fix(document): Correct divergence lifecycle documentation
* fix(document): Neutralize divergence lifecycle prose
* fix(bin): re-arm after an abandoned auto-arm claim and defer a wedge escalation while a worktree is written (#2524)
* fix(watch): re-arm supervision after an abandoned auto-arm claim
A Claude auto-arm cycle that armed, delivered one rewake, and exited left
its single-flight lock behind. Both Stop-event participants then deferred
to that lock forever, because its recorded pid was still live: the
turn-end guard read it as recovery under way and allowed the stop, and the
next Stop firing treated it as another owner and declined to arm. On
2026-08-14 a home with two tasks in flight lost supervision for about 40
minutes with no watcher process and no watcher lock, its beacon frozen at
the one delivery, and both crewmates' finished reports sat in the durable
queue until an operator drained it by hand.
Abandonment is now proven from the epoch ledger instead of inferred from
pid liveness. A lock whose holder pid matches the ledger's own owner_pid
while the recorded outcome is anything other than arming has already
finished its decision, so that claim is reclaimed under the lock's steal
mutex, stops counting as recovery ownership in the guard, and is cleared
by the guard's terminal check rather than deferred to. A failed clear
re-blocks instead of allowing a blind stop, and an arming entry stays in
flight however old it is, because its owner foregrounds the arm for the
whole watcher cycle.
Issue #2251's PR #2263 does not cover this failure. It is closed and
unmerged, lives entirely in bin/fm-watch-arm.sh, and retires the stalled
watcher and matching stale watcher lock of an arm that is currently
running. Here no arm and no watcher were running and no watcher lock
existed, so it has nothing to retire and the home stays blind.
tests/fm-claude-stop-autoarm.test.sh covers the reclaim, the still-arming
and unnamed-owner cases that must keep the gate closed, and the failed
clear. tests/fm-turnend-guard.test.sh covers the guard side of the same
boundary. Both fail without this change.
* fix(watch): defer a wedge escalation while the task worktree is written
The wedge detector had two inputs, rendered pane quietness and the run
step, and neither can see a crew that is writing source, then tests, then
documentation behind a static pane. On 2026-08-14 one crewmate produced
eight consecutive possible-wedge escalations in a single afternoon, three
of them demanding deep inspection, while it was demonstrably working and
then committed. Every one of them cost a supervision turn to disprove by
hand.
Add write activity inside the crew's own recorded worktree as a third
liveness input. crew_worktree_written_since compares the worktree against
the caller's existing idle-window timer file, so -newer needs no clock
arithmetic, no temp file, and no portable mtime write. The probe runs only
inside the branch that was about to escalate, which bounds it to one
pruned, depth-bounded walk per window per FM_STALE_ESCALATE_SECS and
leaves the per-poll stale sweep exactly as cheap as before.
Positive evidence defers rather than cancels. The idle timer restarts so
the next window probes again, the escalation counter is neither advanced
nor reset so a later genuine wedge keeps the demand-deep-inspection
history it earned, and a .writing-since marker ages the whole deferral
chain so the pane still re-surfaces once per FM_PAUSE_RESURFACE_SECS,
through the same throttle shape a declared pause already uses, labeled as
a recheck rather than a wedge. This can only reduce false positives: every
absence of evidence, including no recorded worktree, a torn-down worktree,
a missing anchor, and a failed walk, falls through to the unchanged
escalation schedule, so a crew that writes nothing still escalates on the
existing timetable.
What the signal cannot see, by design or by construction:
- CPU burn with no writes, such as a long compaction, is invisible. That
case keeps the old behavior exactly.
- A commit-only phase writes only .git, which is pruned first so that
firstmate's own read-only git commands against the worktree can never
make the probe self-fulfilling.
- Writes under the pruned generated trees, or deeper than
FM_WORKTREE_WRITE_MAXDEPTH, do not count.
- The probe cannot attribute a write to the crew, so a background build or
another process touching the tree looks the same. The hourly re-surface
is what bounds that, and a churny file cannot buy silence.
- The away-mode daemon's own escalation path is deliberately untouched.
tests/fm-watch-triage.test.sh covers the classifier including the .git
prune, both halves of the live case on one fixture (quiet plus writing
defers, quiet plus silent still escalates and counts), and the bounded
re-surface. All three fail without this change.
* no-mistakes(review): prove autoarm claims by identity; skip mate-home write probe
* no-mistakes(document): document away-mode wedge boundary and probe filesystem limit
* no-mistakes(document): qualify turn-end recovery condition for abandoned auto-arm claims
* fix(watch): keep a write deferral scoped to its own idle window
Two consistency gaps in the worktree write probe, both found while reviewing
the wedge-deferral change on this branch.
A write deferral is a bounded chain: its .writing-since marker ages the whole
chain so a churning worktree still re-surfaces once per resurface window. That
is only sound while the chain belongs to the current quiet stretch, so every
path that restarts the idle-window timer has to drop it too. Two did not: the
corrupt-timer repair in wedge_timer_check, and both first-sight branches for a
captain-relevant status. A chain left over from an earlier quiet stretch made
the first deferral of the new window re-surface immediately instead of after a
full fresh window.
FM_WORKTREE_WRITE_PRUNE is a skip list, so clearing it reads as "skip nothing"
and is the obvious way to widen the probe to the whole depth-bounded tree.
Instead an empty list reported no evidence at all, quietly costing the wedge
detector its third liveness input on a home that meant to widen the walk. An
empty list now widens the walk, and the header says so.
Neither change alters when a stall that writes nothing escalates.
Regressions in tests/fm-watch-triage.test.sh cover all three paths and each
one fails on the pre-fix code.
* no-mistakes(review): honor an empty write-prune, bound the probe, share window_key
* no-mistakes(document): align probe knob count and guard regression-coverage ownership
* no-mistakes(lint): silence deliberate single-quote SC2016 in write-prune env test
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause (#2748)
* fix(bin): give a captain hold the same bounded pause cadence as a declared pause
Two supervisors read a finished task's last status line and disagreed about which
declarations mean an idle endpoint is expected. bin/fm-inactive-reconcile.sh
suppresses its inactive-outcome scan only on `captain-held`, while the away-mode
daemon's wedge path gated deferral on `paused` alone. Both read the LAST line, so
the two verbs are mutually exclusive and no finished task waiting on a person
could satisfy both at once. Marking 11 such tasks `captain-held:` silenced the
900s outcome scan and immediately produced five possible-wedge escalations in one
batch, because the 240s wedge detector no longer saw a pause verb.
fm-classify-lib.sh's status_is_paused_or_captain_held already owns the combined
question, and bin/fm-watch.sh's ordinary-crew wedge path already asked it. This
extends that same answer to the paths still asking the narrower one:
- bin/fm-supervise-daemon.sh, all six sites, which form one subsystem and have to
move together. classify_stale returns the pause action, reconcile_pause_tracking
and migrate_watcher_pause_markers record and migrate the marker, and
housekeeping defers the wedge and then re-surfaces the recheck. Changing only
the stale-persistence gate would defer the escalation while
reconcile_pause_tracking recorded nothing, so the wedge marker would persist and
the sweep would `continue` past it forever: quiet, but never re-surfacing.
- bin/fm-watch.sh's secondmate stale gate, whose downstream owner
pause_state_class already treats both declarations identically.
- bin/fm-push-transition-lib.sh's absorb, where either declaration already names
the human the transition would report and the wait is already durably recorded.
Quieting alone would be half a fix, so the bounded re-surface had to reach a hold
too. A hold has no current-state mapping, unlike `paused`, so authoritative crew
state reports it as unknown and pause_state_class received `none`. An ordinary
crew recovers pause classification from that state through confirmed agent death,
which proves no live decision gate is being silenced. A secondmate's endpoint
liveness is deliberately never read there, because an idle mate is healthy by
design, so that confirmation is unavailable by construction and cannot be
required: without recovering the classification for a mate, every caller silenced
a held mate outright and its hold would rot invisibly. That promotion is bounded
by the declared-wait guard at the top of the function, so it can only reclassify a
task that already declared a wait and shows no positive working evidence.
Two narrow `status_is_paused` calls are deliberately left alone.
bin/fm-crew-state.sh's map_log_state is a current-state reporting contract, not a
wedge path; reporting a hold as `paused` would erase the distinction
status_key_closing_verb and fm-captain-hold.sh depend on, where a `captain-held`
close is a verified durable transfer and a `resolved` close claims outright
settlement. fm-classify-lib.sh's call inside status_is_captain_relevant needs no
change because that function's own case list already returns non-relevant for
`captain-held`.
bin/fm-inactive-reconcile.sh keeps its `captain-held` suppression as it is. Its
guard exists because a finished task's crew state still reports done from a
higher-priority source than the log, and a declared pause needs no such guard: the
scan only reports done or failed, and nothing else reaches its record path.
Widening it would change a separate subsystem's reporting contract, which this
defect does not require.
Coverage extends the existing colocated patterns for these predicates and asserts
both halves. tests/fm-daemon.test.sh covers the classification, the wedge marker
converting to pause tracking with no escalation, the bounded re-surface with its
window reset, and the boundary case where an answered hold stops claiming the
cadence. tests/fm-watch-triage.test.sh covers a held secondmate re-surfacing on
the same bounded cadence without being labeled a wedge.
tests/fm-supervision-events.test.sh covers the absorbed push transition. Every one
of these fails on the pre-fix code except the answered-hold boundary case, which
is there to pin that the quieting was not widened too far.
The `paused:` workaround appended to those 11 tasks is live supervision state and
is untouched here. It can be retired once this lands.
* no-mistakes(review): name the captain in a held task's bounded recheck
* no-mistakes(document): extend declared-wait supervision docs to captain-held holds
* fix(bin): make lint prerequisites and harness tests reliable (#2758)
* fix(lint): name the installer when ShellCheck or actionlint is missing
A missing actionlint exited 127 like a bare command-not-found. Fail with
exit 1 and point at the pinned installer, matching the missing-ShellCheck
path, without weakening the version pin.
* test: isolate kimi and muse detection from inherited Cursor markers
Harness detection checks CURSOR_AGENT before ancestry, so these
markerless-adapter cases failed when the suite itself ran under Cursor.
Clear the verified markers the same way the secondmate harness tests already do.
* no-mistakes(document): Document Muse Cursor marker cleanup
* feat(bin): report watched tooling updates that are available or installed but inert (#2684)
* feat(checks): report tool updates that are available or installed but inert
Firstmate had no way to notice that tooling this home depends on needs an
update, and no way at all to notice the worse case: an update that installed
correctly and then did nothing.
That second case is why this exists. A tool that self-installs into
~/.local/bin while a version manager keeps its own older copy earlier on PATH
looks completely up to date to anything that asks only "is a newer version
published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0
copy stayed earlier on PATH, so every Herdr command failed on a protocol
mismatch and firstmate could not read its own fleet.
bin/fm-tool-update-check.sh reports the two conditions separately:
<tool> update available a newer version exists at the update source.
<tool> update not in effect a newer copy is installed on this host, but
PATH still resolves an older one.
PATH skew is measured, never inferred. Every executable copy of a watched
command on PATH is asked for its own version and those answers are compared,
so one lookup cannot hide the skew, and a directory name is never read as a
version because a version manager's "latest" directory can hold an older
build. A copy that will not report a version is a check failure, not a pass.
The watched tools live in local, gitignored config/watched-tools.json, so
adding a tool is a config edit rather than a code change, and the file is
never propagated to another home. Update sources cover both shapes: a local
clone's commit distance from its remote branch, and a command's own version
and update announcement, including a tool like no-mistakes that prints its
version on one command and announces a new release on another.
The check prints one line when something needs attention and prints nothing
otherwise, so it rides the existing watcher state-check contract with its
trust binding instead of introducing a schedule of its own, and
state/.tool-updates keeps the same pending update from being reported on
every poll.
The check only reports. It never installs, updates, reorders PATH, touches a
version manager, or fetches into a watched repository; every git probe is
read-only.
Tests cover the skew case as a regression, and it was verified by mutation:
removing the skew report, or stopping after the first PATH hit as a single
lookup would, each make that test fail.
* no-mistakes(review): fix tool update check probe reporting, budget, and shim write
* no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets
* no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe
* no-mistakes(review): guard git probes at the budget, record uncut findings
* no-mistakes(document): fix stale watched-tool report-record wording in docs and header
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
The behavior shard's watch-triage suite failed on the new worktree-write wedge
tests. Those five tests are the only ones in the file that do not use its
standard waits. They give a fixed 3 second liveness budget to the one poll that
now spawns the bounded worktree walk, and 4 seconds to an escalating watcher
where every other test in the file gives 10. On a loaded runner that poll
outlives the fixed budget, so the round is reaped before the deferral it asserts
on is recorded, and the test reports a lost deferral instead of the deferral
under test. Wait for a completed poll cycle through the file's own
wait_poll_cycle, which is what its header documents this hazard for, and use the
file's standard 100 tick exit budget.
Verified against a load that reproduces the failure: 11 of 12 runs failed
before, 8 of 8 pass after. Verified by mutation too, so the waits still prove
the behavior: removing the write deferral, and keeping a finished deferral chain
across an idle-timer repair, each still fail their test.
* fix: decouple ask-user decisions from yolo (#2764)
* fix: treat yolo as merge authority only, not ask-user finding authority
Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation.
* no-mistakes(document): Clarify yolo documentation ownership and merge posture
* feat(bin): add a spoken interface that answers from records and hands work over (#2767)
* feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost
Step one of the spoken interface: the laptop captures and plays audio, this
desktop holds the model session, and no AWS credential leaves the desktop.
Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte
of reply audio, 6 runs each, all answered, on a question that forces a records
read:
relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438
direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220
The relay costs about 0.22s of the median. The direct figure reproduces the
earlier survey, which is what makes it a usable control. Excluded: the
captain's own ssh round trip, microphone capture, and speaker output. This
desktop has no microphone and no speaker, so every run used audio files.
Three pieces:
bin/fm-voice-relay.py holds the conversation on this host
bin/fm_voice_records.py what a spoken answer may read, and the handover
bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED
bin/fm_voice_frame.py the wire format both machines share
Real work is handed to the existing bin/fm-inbox.sh rather than a second
queueing surface, and the agent says it is handing over rather than answering
as firstmate.
Read scope: Done history and free-form note bodies are never assembled at any
scope, so the wide default cannot reach the places commercial detail
accumulates. config/voice-read-scope narrows it to counts only, and
config/voice-read-deny excludes a named item in one line. The boundary is an
executable test that widening the reader fails.
Push to talk is the default because it is cheaper and the choice is still open;
--listen open-mic is the single flip.
Two traps worth knowing: a clip with no trailing silence is never answered, and
the end of a reply is contentEnd with stopReason END_TURN, not completionEnd.
A second user turn in one session is treated as barge-in unconditionally, and
an interrupted turn that calls a tool is lost, so the session reconnects per
turn and gives up conversational memory. That is the concrete thing step three
has to solve.
* no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing
* no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir
* no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials
* no-mistakes(review): read account config from home, harden deny and turn failures
* no-mistakes(review): close status verb set, fix inbox help, pair data override
* no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion
* no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline
* no-mistakes(review): survive reader failures, release devices, fix claims
A failure while handling a model event, or while sending a tool result,
left the reader task dead with ended and turn_done clear, and close()
re-raised the stored failure on every await. One dropped stream became a
relay that could never build another session. The reader now reports the
session over in a finally whatever killed it, and close() absorbs the
task the same way it already absorbed its sends.
The laptop client releases what it already started when a later startup
step refuses, SystemExit from the handshake wait included, and names a
device refusal instead of leaking a raw PortAudio error. Whether it
releases correctly against a real device is still unverified here.
The records docstring claimed every reading was filtered to open ids.
Only the pull request count and list are; the worker count and the state
histogram cover every live runtime record, finished ids included,
because a meta file still on disk still needs tearing down.
The finished-work deny half of the suite asserted things that held with
the deny list absent. It is replaced by a deny on an open title, which
removes the row and says so while the count stays honest.
* no-mistakes(review): name reader failures, split file and device refusals
A failure inside the model reader released the waiting turn and told
nobody. The session was not marked spent, no notice reached the client,
and the client waits for a reply end or a notice, so the captain got
their whole timeout of silence and then a record saying the turn went
unanswered with nothing about why. Both ends of the relay now name a
failed turn through one function, once per turn, and --self-test carries
the cause in relay_error the way the client's own record does.
Two things that are not failures stay that way. A stream that simply
ends is the end of a session, which serve still reads on its own terms.
A stream that goes away because close() asked it to is an ordinary
renew, and announcing it would have put a failure notice in front of the
captain on every turn.
On the laptop end, the refusal that became a device error covered the
file-backed playback and capture too, so a mistyped --in-file was
reported as an audio device failure and the advice named the flag that
had just failed. The file ends now report the path and the flag that
chose it and stay an OSError; the device ends keep the device advice and
name the flag for that end. The device paths remain unrun here, so only
the file halves are covered by a test.
* no-mistakes(test): survive model session end, order client turn frames
* no-mistakes(document): sync voice relay docs with reviewed relay behavior
* no-mistakes(document): re-measure relay latency and correct its cause
* no-mistakes(document): correct measurement date and name the unmeasured SSH hop
* no-mistakes(document): describe the unpublished control measurement, fix list formatting
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* fix(bin): preserve Relay follow-up loops until explicit disposition (#2763)
* fix: keep Relay public loops open until retire
Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close.
* no-mistakes(review): Propagate public follow-up registration removal failures
* no-mistakes(review): Persist retire receipts and align parent resolution
* no-mistakes(review): Make rechain resumable after partial obligation creation
* no-mistakes(review): Repair follow-up state, briefs, and expiry escalation
* no-mistakes(review): Serialize follow-up delivery stamps with retirement
* no-mistakes(review): Serialize rechain claims and protect registration terminal states
* no-mistakes(review): Avoid reporting retired delivery loops as open
* no-mistakes(document): Refresh public-loop documentation and verification evidence
* no-mistakes: apply CI fixes
* no-mistakes(review): Preserve delivered follow-up bindings during registration replay
* no-mistakes(review): Harden public follow-up retirement and rechain races
* no-mistakes(review): Fail closed on unresolved secondmate retirement
* no-mistakes(review): Bind secondmate cleanup to its recorded canonical home
* no-mistakes(review): Fix rechain command output and expiry validation
* no-mistakes(review): Validate brief keys and warn on remote promotion
* no-mistakes(document): Document retained public follow-up loops
* no-mistakes(lint): Remove unused bounded-wait loop variable
* feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779)
* feat(bin): merge GitLab merge requests through the guarded PR merge path
bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher,
but bin/fm-pr-merge.sh refused every non-github provider, so a merge request
had to be merged by hand and got none of the recording, guards, or audit
trail a pull request gets.
The merge path now dispatches on the parsed provider. A GitHub URL keeps its
exact previous behavior. A GitLab URL is addressed through glab by the project
URL rebuilt from the parsed host and path, so a merge request on any instance
resolves and no host is hardcoded, and no merge-method flag is added because
the project's own merge method is what should apply.
A GitLab merge happens only after one live read of the merge request confirms
it is open, detailed_merge_status is mergeable, has_conflicts is false,
blocking_discussions_resolved is true, and the head pipeline succeeded at the
exact current head. Every failing condition is reported, not just the first.
The verified head is bound to the merge with glab's --sha, so a push landing
between the read and the merge fails the merge instead of landing commits
nothing verified. Recorded metadata is never the authority for any of this: a
rebase moves the head and leaves a recorded value stale, so a recorded head
that disagrees with the live one is reported rather than trusted, and the
recorded value is read before the recording step because that step drops a
GitLab head it cannot resolve.
* no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent
* no-mistakes(test): state authorised GitHub narrowing of bundled -R guard
This branch NARROWS GitHub behaviour. The narrowing was authorised
deliberately rather than slipping in by accident, and it applies to both
providers, GitHub and GitLab alike, because a script that guards one provider
and not the other is a trap for the next reader.
What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a
bundled short-option cluster that includes R, for example "-dR other/repo".
The forge CLIs expand such a cluster one character at a time, so it carries
"--repo other/repo", and that later value wins over the repository the URL
named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR
other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR
other/repo" and exited 0 with pr= recorded and the merge poll armed. It now
exits 1 with "extra merge arguments must not override the repository", records
nothing, and invokes no forge merge command. Every other GitHub invocation is
byte-identical to the base commit.
Closing that hole honours the existing rule rather than departing from it. The
file header already forbids --repo and -R because the repository must come
only from the URL, so a bundled cluster carrying a repository override was
never legitimate behaviour to preserve: it was that guard being evaded.
Redirecting a merge to a repository the URL does not name is exactly what the
guard exists to prevent.
The refusal is already pinned on both paths by the existing case
test_bundled_repo_override_args_refuse_before_recording in
tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR
https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in
the task meta, no armed merge poll, and no forge merge command invoked, with a
control case proving a cluster that carries no repository override still
reaches the forge. No duplicate assertion was added. Both assertions were
confirmed to have teeth by narrowing the guard back to a bare -R and watching
each path fail.
This commit carries no file change: the guard and its coverage landed in
614853d, and this message exists so the pull request description states the
narrowing.
* no-mistakes(document): fix README pointer for GitLab watch and merge doc
* no-mistakes: apply CI fixes
* fix(bin): record a lost relay connection instead of an unanswered turn (#2788)
* no-mistakes: apply CI fixes
* fix(bin): drop a private record citation and narrow the review rule
Three corrections to the spoken interface that landed in #2767, plus one
fix carried over from that branch after its pull request had already been
merged.
The confidentiality fix. The module docstring of bin/fm-voice-relay.py
cited a private, gitignored fleet record by exact path and section number.
That widens what this public repository points at, and it cannot resolve
for any reader here, because the path has never been in the repository.
Both traps it pointed at are already described in full in the list
immediately below it, and docs/voice-relay.md carries the same two for
operators with no citation at all, so the pointer is removed and no claim
is weakened by losing it. Two comments that referred to "the survey" as
though it were something a reader could open are reworded the same way.
Neither exposed a path, so that half is comprehensibility rather than
confidentiality.
The review rule. .greptile/rules.md is kept, because its conditions are
right and deleting it would leave the next reviewer to re-litigate a
decision already argued out. What was wrong with it is narrower than its
existence: it read as settled repository policy, when whether VISION.md
itself should be reconciled is an open question belonging to the captain.
One sentence now says so, and says that the conditions listed below it are
what the interpretation depends on. That narrows the claim rather than
widening it.
The carried-over fix. The first commit on this branch is 7f98e797 from
fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes
the window where a transport failure was recorded and then erased, so a
run could be emitted as answered false with relay_error null. That matters
more than it looks: relay_error is the field that keeps an infrastructure
failure from being averaged into a latency figure, so the failure mode is
a dead connection wearing the costume of a slow reply. It landed fifteen
minutes after #2767 merged and so never reached the default branch.
* no-mistakes(review): name a reason on every unanswered-turn close path
* no-mistakes(review): guard the downlink body and pin frames to their turn
* no-mistakes(review): attribute reply audio to its own turn and tell endings apart
* no-mistakes(review): tell a cut-short reply from an unanswered turn
* no-mistakes(review): discard reply audio arriving after the output closes
* no-mistakes(review): count discarded reply audio on the speaker path too
* no-mistakes(review): keep a reason off a turn already answered in full
* no-mistakes(review): say a reset cut a reply short, not that none arrived
* no-mistakes(review): read one turn's audio count once, and hush a tidy exit
* no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide
* fix(composer): stop a blocked pi pane from proving an empty composer (#2811)
A pi worker parked on an interactive prompt - a permission dialog, a
question menu, a trust dialog - reports agent_status=blocked, because it
is waiting on a human keystroke. Pi draws that menu above its separator
pair, so the composer region between the rules is blank and structure
alone looks like a free composer. _fm_composer_pi_verdict admitted
blocked alongside idle and done, so the shared classifier reported an
affirmatively empty composer for exactly the pane where typing is unsafe.
Every "is it safe to type here?" consumer reads that verdict and proceeds
only on an affirmative empty, so both are told yes on a parked prompt:
the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's
pre-type refusal. The keys then answer the menu instead of composing a
message - the highlighted default is selected, the text is discarded, and
the record attributes a decision to a human who never made it.
blocked now defers to unknown, which every consumer already treats as
fail-closed. idle and done still prove an empty composer, so ordinary
steering is unchanged, and Cursor is unaffected because its always-blocked
panes never reach this pi-only branch.
Regression coverage lands first at both levels: the verdict owner
(a blocked pi defers) and the herdr adapter (a parked pi prompt is not an
empty composer).
* fix(bin): require project clone roots during fleet sync (#2849)
* fix(bin): require a clone root before fleet-sync touches a project
Git repository discovery walks upward, so `git -C projects/<dir>` on a plain
directory nested under projects/ resolves to the enclosing repository - in a
firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its
candidates with `rev-parse --is-inside-work-tree`, which such a directory
passes, so every later git call read, pruned and fast-forwarded firstmate's own
default branch and reported it under the project directory's label. A running
session's AGENTS.md changed underneath it, and the report named a project that
had nothing to do with the change.
Require each candidate to be the root of its own work tree before any other git
command: compare `rev-parse --show-toplevel` against the directory's own
physical path. Both sides are physical, so a symlinked clone still compares
equal. Anything else is skipped by name, naming the repository that would have
been touched, and bootstrap relays that as a FLEET_SYNC line.
Regression coverage reproduces the wrong-repo fast-forward against a home nested
inside another repository, in both the whole-fleet and single-project forms, and
pins that a symlinked clone dir still syncs.
* no-mistakes(review): Keep enclosing fixture clean during clone-root regression
* fix(bin): retry transient Lavish poll interruptions (#2846)
* fix(procevent): retry a transient Lavish poll interruption quietly
A live Lavish listener can be cut short by the server with exactly
error: Lavish Editor poll response was interrupted
code: SERVER_ERROR
while the session's marks remain available. Firstmate registered raw
`lavish-axi poll` output, so the generic process-event runner captured
that transient response as a result and woke the whole fleet over what is
really an internal retry.
The Lavish adapter now registers its own listener command, which reruns
the published blocking poll up to 12 times at 5 second intervals for that
one exact two-line response. The match is deliberately narrow: real
feedback, ended and missing sessions, any other SERVER_ERROR, and the same
interruption still standing once the bound is spent all pass straight
through and are captured and announced as before. The retry is a Lavish
fact, so the generic runner stays adapter-agnostic.
`FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the
interval only, refused rather than rounded when malformed, so a test can
exercise the real bound without waiting it out.
* no-mistakes(review): Harden Lavish retry matching, validation, and cleanup
* no-mistakes(review): Bound Lavish retry staging and stabilize regression
* no-mistakes(document): docs: explain Lavish retry adoption
* no-mistakes(lint): Restore Lavish trap ShellCheck suppression
* fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838)
The unguarded Herdr declaration quoted `{TASK}` in its own prose while the
scaffold instructs firstmate to replace every `{TASK}` placeholder. The
documented global replace therefore spliced the whole task body into the
middle of the safety gate's sentence, silently destroying the one contract
that exists precisely because the scaffold cannot inspect the task text.
Reword the gate to refer to the task text filled in above, leaving the
placeholder only at its genuine fill site. Rewording rather than renaming the
token keeps the unfilled-charter guards in fm-home-seed.sh and
fm-remote-home-seed.sh working unchanged.
Add a regression test that performs the documented global fill on ship and
scout scaffolds and asserts the body lands once and the gate survives.
* fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837)
The writer lock's stale-lock branch read the lock's mtime with
`stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem
stat, so it consumed the format string as a path, complained on stderr,
printed a partial filesystem dump (" File: ...") on stdout, and still
exited 0. The GNU form in the fallback therefore never ran, and the
following arithmetic evaluated the word `File`, aborting the writer under
`set -u` with "File: unbound variable".
fm-teardown.sh died there after returning the worktree, leaving
state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and
.turn-ended behind. The surviving metadata kept the watcher monitoring an
endpoint whose agent was gone, so a finished task produced stale wakes
forever, and every re-run died identically because the abandoned lock was
never broken.
Detect the platform once and pick the right stat form, the pattern
bin/fm-watch.sh already documents, and treat any non-numeric result as
"just created" so a future portability surprise degrades to a lock-timeout
refusal rather than killing teardown mid-way.
* fix(stow): add opt-in pass horizon for memory decay (#2850)
* fix(stow): give memory decay a per-pass horizon so the clock fires
The tiered decay clocks were wall-clock only, while admission is per-pass:
each /stow admits the findings that pass produced. In a home that stows
daily those two rates diverge by the stow cadence, an entry the fleet keeps
exercising never reaches 30 days unreinforced, and memory only grows while
the pass reports decay evaluated.
Give each dated marker an optional unreinforced-pass counter and make both
tiers stale at whichever horizon comes first: 10 passes or 30 days for
aging, 3 passes or 7 days for perishable. Reinforcement clears the counter
and nothing else does, so the existing evidence-based restamp rule stays
the only way an entry renews its lease. An absent /N means zero, so entries
that stay exercised carry no extra marker bytes, and a rarely stowed home
keeps its current behaviour through the unchanged date horizon.
* no-mistakes(document): Align stow workflow with dual decay clocks
* fix(stow): make the per-pass decay horizon opt-in
The unreinforced-pass horizon shipped as a new default archival cadence,
which is a product default rather than a restoration of the existing
wall-clock contract. Keep the 30-day and 7-day horizons as the only
default clock, and put the 10-pass and 3-pass horizons behind an explicit
opt-in: config/stow-pass-horizon for the firstmate home, and the file's
own header pointer for the public skill.
With the opt-in absent no counter is written and no counter is read, so a
home that does not ask for it decays exactly as it does today.
* no-mistakes(review): Preserve frozen counters and correct archive provenance
* test(watcher): stop fixture confirmation budgets racing real child startup (#2876)
tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently
under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its
confirmation deadline immediately after forking the real child watcher, so the
child's entire fork, exec, lock acquisition and beacon publication has to land
inside that wall clock. Two cases shrank that budget to one second, leaving a
two-second window for work measured at 3.1-4.9s under CPU oversubscription, so
the arm honestly reported "FAILED - no live watcher with a fresh beacon" and
their premises collapsed. A third case ran on the production budget, but its
child must also execute a registered check before exiting: measured at 1.9-2.3s
idle and 9.1-13.1s under load, against an 11s budget.
The two cases that must confirm a real child now hold the arm to production's
own budget instead of a shrunken fixture one, the immediate-wake case gets an
explicit budget with headroom over its measured loaded cost, and the two waits
for the arm's typed failure are sized off the largest production default rather
than a fixed eight seconds.
No bin/ change and no default behavior change: the lock's fail-closed semantics,
SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are
untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after
3/3 red before the change, and CONTRIBUTING.md records the convention.
* fix(bin): deterministically order remote tool paths (#2870)
* fix(bin): order discovered tool installs by the shell's own expansion
fm_remote_job_compose_operator_path built the asdf and mise install
directories with `compgen -G`, which does not sort. Bash sorts glob
matches in pathexp.c, on the shell's own pathname-expansion path only;
`compgen -G` reaches the same glob_filename through pcomplete.c, which
sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3
that handed the composition raw readdir order, so which install of a
multi-version tool a remote job resolved was decided by directory order
on disk rather than by this composition.
Expand the globs at the call sites and let the function take the matches,
so the composition and the documented portable-PATH contract are the same
operation. Quoting the account home at the call site also stops a home
whose name contains glob metacharacters from being reinterpreted.
The colocated regression pins both the order and the mechanism: bash 5.3
moved sorting into the glob library, so an order-only assertion cannot
see the defect there.
* no-mistakes(review): Remove source-reading PATH regression guard
* fix(bin): prevent routed secondmate work from stranding (#2848)
* fix: surface stalled secondmate queues and wake handoffs
* no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe
* no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery
* no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent
* no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation
* no-mistakes(review): Reconcile correlated handoff wake delivery after crashes
* no-mistakes(review): Keep failed wakes retryable and isolate stall receipts
* no-mistakes(review): Reset known-undelivered wake attempts for durable retries
* no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts
* no-mistakes(review): Atomically restore retryability after reconciled send failures
* no-mistakes(review): Serialize delivery confirmation with reconciliation
* no-mistakes(document): Document routed wake and stall supervision
* no-mistakes(lint): Fix ShellCheck expansion and subshell warnings
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Retire stale wake state and defer pre-move wakes
* no-mistakes(review): Secure markers, bind batches, and preserve teardown routes
* no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs
* no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs
* no-mistakes(document): Document prepared wake batch ownership
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* no-mistakes(review): Make local wake retirement recoverable
* no-mistakes(document): Clarify handoff recovery and teardown documentation
* fix: make macOS inbox test path portable (#2857)
* feat(bin): deliver local steers through durable task inboxes (#2856)
* feat(bin): steer local tasks by durable inbox record plus constant doorbell
Stage 1 (local steers) of the captain-adopted reframe in
data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text
steer to a task recorded in this home is appended as a sequenced durable
record under state/<id>.inbox/ and the terminal receives only one constant
self-describing doorbell line, best-effort. The worker acknowledges by
moving the record into handled/; the watcher re-rings an unacknowledged
message on an idle pane and escalates once as an ordinary stale wake.
--resolve-key closes decisions at enqueue time, because the durable
enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns
the record format, doorbell line, and re-ring ladder.
The typed plane remains for what must reach the terminal itself:
lifecycle keys, harness-native slash and codex $-skill invocations,
explicit backend targets, and the remote secondmate leg (unchanged until
the remote inbox leg ships separately). The composer classifier is
demoted from delivery proof to an advisory ring guard that skips only on
a proven pending verdict.
Verified live against claude, codex, opencode, pi, grok, and muse: each
real worker read its record, acted, and acked with the mv
(docs/verification/runtime-backends.md "Steering-inbox doorbell").
* docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run
* test(captain-hold): read the chat-channel answer from the durable inbox record
* test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite
* no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements
* no-mistakes(review): Serialize watcher actions with inbox acknowledgements
* no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks
* no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery
* no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown
* no-mistakes(document): Update inbox and typed-plane documentation
* no-mistakes: apply CI fixes
* no-mistakes: apply CI fixes
* revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1
The CI monitor's fix changed the secondmate marking contract for
parser-native invocations (appending the marker after the text) and
softened the both-commit-and-marker-failed branch to exit 0. The merge
authority ruled the marking question out of scope for this stage-1
transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and
ruled the both-failed case a loud nonzero local failure. Restore both,
keeping the monitor's legitimate migrations and hardening.
* no-mistakes(document): Document inbox and typed-plane boundaries
* no-mistakes(document): Scope backend transport docs to typed plane
* no-mistakes(document): Clarify inbox attempt-budget documentation
* no-mistakes: apply CI fixes
* fix(send): the durable record alone governs the inbox exit status
Captain-refined ruling on the F2/Greptile finding: the durable inbox
record is what delivers the steer, so pending-reply bookkeeping trouble
after a successful enqueue never exits nonzero - a resend-inviting status
would make automated callers enqueue the delivered instruction again
under a new sequence. With the recovery marker stored the watcher
reconciles silently; with the commit and marker both lost the send
surfaces a distinct reply-tracking-degraded do-not-resend warning and
still exits 0. Nonzero remains only where nothing was delivered (or a
decision close needs its manual command). Regression: record durable +
both bookkeeping writes lost -> exit 0, one record, no duplicate.
* no-mistakes(review): Preserve inbox ordering with drain-all doorbells
* no-mistakes(review): Surface unwritable inbox ladder bookkeeping
* no-mistakes(review): Silence ladder failures after inbox acknowledgement
* no-mistakes(document): Update steering inbox documentation
* no-mistakes: apply CI fixes
* feat(bin): add fast local lint mode (#2891)
* feat: add fast local lint mode
* fix: preserve complete fm-lint help
* fix: isolate fast lint mode
* no-mistakes(document): Clarify lint mode documentation ownership
* no-mistakes: apply CI fixes
* feat(bin): deliver remote steers through durable inboxes (#2901)
* feat(bin): deliver remote secondmate steers through durable task in…
yelenplays
added a commit
to yelenplays/firstmate
that referenced
this pull request
Sep 18, 2026
…in operations (#33) * feat(bin): report watched tooling updates that are available or installed but inert (#2684) * feat(checks): report tool updates that are available or installed but inert Firstmate had no way to notice that tooling this home depends on needs an update, and no way at all to notice the worse case: an update that installed correctly and then did nothing. That second case is why this exists. A tool that self-installs into ~/.local/bin while a version manager keeps its own older copy earlier on PATH looks completely up to date to anything that asks only "is a newer version published". On 2026-08-20 a Herdr update landed at 0.8.2 while an older 0.8.0 copy stayed earlier on PATH, so every Herdr command failed on a protocol mismatch and firstmate could not read its own fleet. bin/fm-tool-update-check.sh reports the two conditions separately: <tool> update available a newer version exists at the update source. <tool> update not in effect a newer copy is installed on this host, but PATH still resolves an older one. PATH skew is measured, never inferred. Every executable copy of a watched command on PATH is asked for its own version and those answers are compared, so one lookup cannot hide the skew, and a directory name is never read as a version because a version manager's "latest" directory can hold an older build. A copy that will not report a version is a check failure, not a pass. The watched tools live in local, gitignored config/watched-tools.json, so adding a tool is a config edit rather than a code change, and the file is never propagated to another home. Update sources cover both shapes: a local clone's commit distance from its remote branch, and a command's own version and update announcement, including a tool like no-mistakes that prints its version on one command and announces a new release on another. The check prints one line when something needs attention and prints nothing otherwise, so it rides the existing watcher state-check contract with its trust binding instead of introducing a schedule of its own, and state/.tool-updates keeps the same pending update from being reported on every poll. The check only reports. It never installs, updates, reorders PATH, touches a version manager, or fetches into a watched repository; every git probe is read-only. Tests cover the skew case as a regression, and it was verified by mutation: removing the skew report, or stopping after the first PATH hit as a single lookup would, each make that test fail. * no-mistakes(review): fix tool update check probe reporting, budget, and shim write * no-mistakes(review): keep sweeps alive on broken patterns and oversized budgets * no-mistakes(review): roll back failed arm, widen budget clamp, bound repo probe * no-mistakes(review): guard git probes at the budget, record uncut findings * no-mistakes(document): fix stale watched-tool report-record wording in docs and header * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes The behavior shard's watch-triage suite failed on the new worktree-write wedge tests. Those five tests are the only ones in the file that do not use its standard waits. They give a fixed 3 second liveness budget to the one poll that now spawns the bounded worktree walk, and 4 seconds to an escalating watcher where every other test in the file gives 10. On a loaded runner that poll outlives the fixed budget, so the round is reaped before the deferral it asserts on is recorded, and the test reports a lost deferral instead of the deferral under test. Wait for a completed poll cycle through the file's own wait_poll_cycle, which is what its header documents this hazard for, and use the file's standard 100 tick exit budget. Verified against a load that reproduces the failure: 11 of 12 runs failed before, 8 of 8 pass after. Verified by mutation too, so the waits still prove the behavior: removing the write deferral, and keeping a finished deferral chain across an idle-timer repair, each still fail their test. * fix: decouple ask-user decisions from yolo (#2764) * fix: treat yolo as merge authority only, not ask-user finding authority Yolo on/off was documented as also deciding no-mistakes ask-user findings, which hid firstmate's duty to judge unambiguous-toward-design findings itself. Keep every safety boundary; this is a contract clarification, not a relaxation. * no-mistakes(document): Clarify yolo documentation ownership and merge posture * feat(bin): add a spoken interface that answers from records and hands work over (#2767) * feat(voice): spoken round trip on Nova Sonic 2 with a measured relay cost Step one of the spoken interface: the laptop captures and plays audio, this desktop holds the model session, and no AWS credential leaves the desktop. Measured, amazon.nova-2-sonic-v1:0 in eu-north-1, end of speech to first byte of reply audio, 6 runs each, all answered, on a question that forces a records read: relay path 1.229 1.379 1.428 1.447 1.481 1.516 median 1.438 direct 1.147 1.179 1.203 1.237 1.244 1.317 median 1.220 The relay costs about 0.22s of the median. The direct figure reproduces the earlier survey, which is what makes it a usable control. Excluded: the captain's own ssh round trip, microphone capture, and speaker output. This desktop has no microphone and no speaker, so every run used audio files. Three pieces: bin/fm-voice-relay.py holds the conversation on this host bin/fm_voice_records.py what a spoken answer may read, and the handover bin/fm-voice-client.py the laptop end; audio devices UNVERIFIED bin/fm_voice_frame.py the wire format both machines share Real work is handed to the existing bin/fm-inbox.sh rather than a second queueing surface, and the agent says it is handing over rather than answering as firstmate. Read scope: Done history and free-form note bodies are never assembled at any scope, so the wide default cannot reach the places commercial detail accumulates. config/voice-read-scope narrows it to counts only, and config/voice-read-deny excludes a named item in one line. The boundary is an executable test that widening the reader fails. Push to talk is the default because it is cheaper and the choice is still open; --listen open-mic is the single flip. Two traps worth knowing: a clip with no trailing silence is never answered, and the end of a reply is contentEnd with stopReason END_TURN, not completionEnd. A second user turn in one session is treated as barge-in unconditionally, and an interrupted turn that calls a tool is lost, so the session reconnects per turn and gives up conversational memory. That is the concrete thing step three has to solve. * no-mistakes(review): fix voice relay credential reuse, frame validation and record parsing * no-mistakes(review): test uplink header guard, bound unknown expiry, align state dir * no-mistakes(review): decide deny per item, guard turn failures, bound ambient credentials * no-mistakes(review): read account config from home, harden deny and turn failures * no-mistakes(review): close status verb set, fix inbox help, pair data override * no-mistakes(review): keep profile-free relay alive, unblock loop, fix dead assertion * no-mistakes(review): hide finished pull requests, refuse open mic, keep suite offline * no-mistakes(review): survive reader failures, release devices, fix claims A failure while handling a model event, or while sending a tool result, left the reader task dead with ended and turn_done clear, and close() re-raised the stored failure on every await. One dropped stream became a relay that could never build another session. The reader now reports the session over in a finally whatever killed it, and close() absorbs the task the same way it already absorbed its sends. The laptop client releases what it already started when a later startup step refuses, SystemExit from the handshake wait included, and names a device refusal instead of leaking a raw PortAudio error. Whether it releases correctly against a real device is still unverified here. The records docstring claimed every reading was filtered to open ids. Only the pull request count and list are; the worker count and the state histogram cover every live runtime record, finished ids included, because a meta file still on disk still needs tearing down. The finished-work deny half of the suite asserted things that held with the deny list absent. It is replaced by a deny on an open title, which removes the row and says so while the count stays honest. * no-mistakes(review): name reader failures, split file and device refusals A failure inside the model reader released the waiting turn and told nobody. The session was not marked spent, no notice reached the client, and the client waits for a reply end or a notice, so the captain got their whole timeout of silence and then a record saying the turn went unanswered with nothing about why. Both ends of the relay now name a failed turn through one function, once per turn, and --self-test carries the cause in relay_error the way the client's own record does. Two things that are not failures stay that way. A stream that simply ends is the end of a session, which serve still reads on its own terms. A stream that goes away because close() asked it to is an ordinary renew, and announcing it would have put a failure notice in front of the captain on every turn. On the laptop end, the refusal that became a device error covered the file-backed playback and capture too, so a mistyped --in-file was reported as an audio device failure and the advice named the flag that had just failed. The file ends now report the path and the flag that chose it and stay an OSError; the device ends keep the device advice and name the flag for that end. The device paths remain unrun here, so only the file halves are covered by a test. * no-mistakes(test): survive model session end, order client turn frames * no-mistakes(document): sync voice relay docs with reviewed relay behavior * no-mistakes(document): re-measure relay latency and correct its cause * no-mistakes(document): correct measurement date and name the unmeasured SSH hop * no-mistakes(document): describe the unpublished control measurement, fix list formatting * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): preserve Relay follow-up loops until explicit disposition (#2763) * fix: keep Relay public loops open until retire Delivering a promised-final reply was deleting the only record that tied a public thread to later work, so a follow-on ship silently owed no closing reply. Retain the registration after delivery, rechain follow-on work onto the same thread, and make retire --reason the only close. * no-mistakes(review): Propagate public follow-up registration removal failures * no-mistakes(review): Persist retire receipts and align parent resolution * no-mistakes(review): Make rechain resumable after partial obligation creation * no-mistakes(review): Repair follow-up state, briefs, and expiry escalation * no-mistakes(review): Serialize follow-up delivery stamps with retirement * no-mistakes(review): Serialize rechain claims and protect registration terminal states * no-mistakes(review): Avoid reporting retired delivery loops as open * no-mistakes(document): Refresh public-loop documentation and verification evidence * no-mistakes: apply CI fixes * no-mistakes(review): Preserve delivered follow-up bindings during registration replay * no-mistakes(review): Harden public follow-up retirement and rechain races * no-mistakes(review): Fail closed on unresolved secondmate retirement * no-mistakes(review): Bind secondmate cleanup to its recorded canonical home * no-mistakes(review): Fix rechain command output and expiry validation * no-mistakes(review): Validate brief keys and warn on remote promotion * no-mistakes(document): Document retained public follow-up loops * no-mistakes(lint): Remove unused bounded-wait loop variable * feat(bin): merge GitLab merge requests through the guarded PR merge path (#2779) * feat(bin): merge GitLab merge requests through the guarded PR merge path bin/fm-pr-lib.sh already parses a GitLab merge request URL for the watcher, but bin/fm-pr-merge.sh refused every non-github provider, so a merge request had to be merged by hand and got none of the recording, guards, or audit trail a pull request gets. The merge path now dispatches on the parsed provider. A GitHub URL keeps its exact previous behavior. A GitLab URL is addressed through glab by the project URL rebuilt from the parsed host and path, so a merge request on any instance resolves and no host is hardcoded, and no merge-method flag is added because the project's own merge method is what should apply. A GitLab merge happens only after one live read of the merge request confirms it is open, detailed_merge_status is mergeable, has_conflicts is false, blocking_discussions_resolved is true, and the head pipeline succeeded at the exact current head. Every failing condition is reported, not just the first. The verified head is bound to the merge with glab's --sha, so a push landing between the read and the merge fails the merge instead of landing commits nothing verified. Recorded metadata is never the authority for any of this: a rebase moves the head and leaves a recorded value stale, so a recorded head that disagrees with the live one is reported rather than trusted, and the recorded value is read before the recording step because that step drops a GitLab head it cannot resolve. * no-mistakes(review): reject bundled -R clusters and make tool-absence cases host-independent * no-mistakes(test): state authorised GitHub narrowing of bundled -R guard This branch NARROWS GitHub behaviour. The narrowing was authorised deliberately rather than slipping in by accident, and it applies to both providers, GitHub and GitLab alike, because a script that guards one provider and not the other is a trap for the next reader. What bin/fm-pr-merge.sh now refuses is extra merge arguments containing a bundled short-option cluster that includes R, for example "-dR other/repo". The forge CLIs expand such a cluster one character at a time, so it carries "--repo other/repo", and that later value wins over the repository the URL named. Before this change, "fm-pr-merge.sh <task> <github-url> -- -dR other/repo" reached "gh-axi pr merge 12 --repo example/repo --squash -dR other/repo" and exited 0 with pr= recorded and the merge poll armed. It now exits 1 with "extra merge arguments must not override the repository", records nothing, and invokes no forge merge command. Every other GitHub invocation is byte-identical to the base commit. Closing that hole honours the existing rule rather than departing from it. The file header already forbids --repo and -R because the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent. The refusal is already pinned on both paths by the existing case test_bundled_repo_override_args_refuse_before_recording in tests/fm-pr-merge.test.sh. On GitHub ("-dR wrong/repo") and on GitLab ("-yR https://other.example/g/p") it asserts exit 1, the refusal wording, no pr= in the task meta, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster that carries no repository override still reaches the forge. No duplicate assertion was added. Both assertions were confirmed to have teeth by narrowing the guard back to a bare -R and watching each path fail. This commit carries no file change: the guard and its coverage landed in 614853d, and this message exists so the pull request description states the narrowing. * no-mistakes(document): fix README pointer for GitLab watch and merge doc * no-mistakes: apply CI fixes * fix(bin): record a lost relay connection instead of an unanswered turn (#2788) * no-mistakes: apply CI fixes * fix(bin): drop a private record citation and narrow the review rule Three corrections to the spoken interface that landed in #2767, plus one fix carried over from that branch after its pull request had already been merged. The confidentiality fix. The module docstring of bin/fm-voice-relay.py cited a private, gitignored fleet record by exact path and section number. That widens what this public repository points at, and it cannot resolve for any reader here, because the path has never been in the repository. Both traps it pointed at are already described in full in the list immediately below it, and docs/voice-relay.md carries the same two for operators with no citation at all, so the pointer is removed and no claim is weakened by losing it. Two comments that referred to "the survey" as though it were something a reader could open are reworded the same way. Neither exposed a path, so that half is comprehensibility rather than confidentiality. The review rule. .greptile/rules.md is kept, because its conditions are right and deleting it would leave the next reviewer to re-litigate a decision already argued out. What was wrong with it is narrower than its existence: it read as settled repository policy, when whether VISION.md itself should be reconciled is an open question belonging to the captain. One sentence now says so, and says that the conditions listed below it are what the interpretation depends on. That narrows the claim rather than widening it. The carried-over fix. The first commit on this branch is 7f98e797 from fm/voice-relay-build-v4, taken verbatim rather than rewritten. It closes the window where a transport failure was recorded and then erased, so a run could be emitted as answered false with relay_error null. That matters more than it looks: relay_error is the field that keeps an infrastructure failure from being averaged into a latency figure, so the failure mode is a dead connection wearing the costume of a slow reply. It landed fifteen minutes after #2767 merged and so never reached the default branch. * no-mistakes(review): name a reason on every unanswered-turn close path * no-mistakes(review): guard the downlink body and pin frames to their turn * no-mistakes(review): attribute reply audio to its own turn and tell endings apart * no-mistakes(review): tell a cut-short reply from an unanswered turn * no-mistakes(review): discard reply audio arriving after the output closes * no-mistakes(review): count discarded reply audio on the speaker path too * no-mistakes(review): keep a reason off a turn already answered in full * no-mistakes(review): say a reset cut a reply short, not that none arrived * no-mistakes(review): read one turn's audio count once, and hush a tidy exit * no-mistakes(document): fix stale session-end relay_error claim in voice-relay guide * fix(composer): stop a blocked pi pane from proving an empty composer (#2811) A pi worker parked on an interactive prompt - a permission dialog, a question menu, a trust dialog - reports agent_status=blocked, because it is waiting on a human keystroke. Pi draws that menu above its separator pair, so the composer region between the rules is blank and structure alone looks like a free composer. _fm_composer_pi_verdict admitted blocked alongside idle and done, so the shared classifier reported an affirmatively empty composer for exactly the pane where typing is unsafe. Every "is it safe to type here?" consumer reads that verdict and proceeds only on an affirmative empty, so both are told yes on a parked prompt: the away-mode injection guard in bin/fm-supervise-daemon.sh, and fm-send's pre-type refusal. The keys then answer the menu instead of composing a message - the highlighted default is selected, the text is discarded, and the record attributes a decision to a human who never made it. blocked now defers to unknown, which every consumer already treats as fail-closed. idle and done still prove an empty composer, so ordinary steering is unchanged, and Cursor is unaffected because its always-blocked panes never reach this pi-only branch. Regression coverage lands first at both levels: the verdict owner (a blocked pi defers) and the herdr adapter (a parked pi prompt is not an empty composer). * fix(bin): require project clone roots during fleet sync (#2849) * fix(bin): require a clone root before fleet-sync touches a project Git repository discovery walks upward, so `git -C projects/<dir>` on a plain directory nested under projects/ resolves to the enclosing repository - in a firstmate home, the firstmate checkout itself. fm-fleet-sync.sh guarded its candidates with `rev-parse --is-inside-work-tree`, which such a directory passes, so every later git call read, pruned and fast-forwarded firstmate's own default branch and reported it under the project directory's label. A running session's AGENTS.md changed underneath it, and the report named a project that had nothing to do with the change. Require each candidate to be the root of its own work tree before any other git command: compare `rev-parse --show-toplevel` against the directory's own physical path. Both sides are physical, so a symlinked clone still compares equal. Anything else is skipped by name, naming the repository that would have been touched, and bootstrap relays that as a FLEET_SYNC line. Regression coverage reproduces the wrong-repo fast-forward against a home nested inside another repository, in both the whole-fleet and single-project forms, and pins that a symlinked clone dir still syncs. * no-mistakes(review): Keep enclosing fixture clean during clone-root regression * fix(bin): retry transient Lavish poll interruptions (#2846) * fix(procevent): retry a transient Lavish poll interruption quietly A live Lavish listener can be cut short by the server with exactly error: Lavish Editor poll response was interrupted code: SERVER_ERROR while the session's marks remain available. Firstmate registered raw `lavish-axi poll` output, so the generic process-event runner captured that transient response as a result and woke the whole fleet over what is really an internal retry. The Lavish adapter now registers its own listener command, which reruns the published blocking poll up to 12 times at 5 second intervals for that one exact two-line response. The match is deliberately narrow: real feedback, ended and missing sessions, any other SERVER_ERROR, and the same interruption still standing once the bound is spent all pass straight through and are captured and announced as before. The retry is a Lavish fact, so the generic runner stays adapter-agnostic. `FM_LAVISH_POLL_RETRY_DELAY` is a bounded 0 to 60 second override for the interval only, refused rather than rounded when malformed, so a test can exercise the real bound without waiting it out. * no-mistakes(review): Harden Lavish retry matching, validation, and cleanup * no-mistakes(review): Bound Lavish retry staging and stabilize regression * no-mistakes(document): docs: explain Lavish retry adoption * no-mistakes(lint): Restore Lavish trap ShellCheck suppression * fix(brief): stop the documented {TASK} fill from corrupting the Herdr gate (#2838) The unguarded Herdr declaration quoted `{TASK}` in its own prose while the scaffold instructs firstmate to replace every `{TASK}` placeholder. The documented global replace therefore spliced the whole task body into the middle of the safety gate's sentence, silently destroying the one contract that exists precisely because the scaffold cannot inspect the task text. Reword the gate to refer to the task text filled in above, leaving the placeholder only at its genuine fill site. Rewording rather than renaming the token keeps the unfilled-charter guards in fm-home-seed.sh and fm-remote-home-seed.sh working unchanged. Add a regression test that performs the documented global fill on ship and scout scaffolds and asserts the body lands once and the gate survives. * fix(bin): resolve the busy-state lock mtime with the platform's own stat form (#2837) The writer lock's stale-lock branch read the lock's mtime with `stat -f %m ... || stat -c %Y ...`. On GNU coreutils `-f` is filesystem stat, so it consumed the format string as a path, complained on stderr, printed a partial filesystem dump (" File: ...") on stdout, and still exited 0. The GNU form in the fallback therefore never ran, and the following arithmetic evaluated the word `File`, aborting the writer under `set -u` with "File: unbound variable". fm-teardown.sh died there after returning the worktree, leaving state/<id>.meta, .status, .busy-gen, .busy-state, .busy-state.lock/ and .turn-ended behind. The surviving metadata kept the watcher monitoring an endpoint whose agent was gone, so a finished task produced stale wakes forever, and every re-run died identically because the abandoned lock was never broken. Detect the platform once and pick the right stat form, the pattern bin/fm-watch.sh already documents, and treat any non-numeric result as "just created" so a future portability surprise degrades to a lock-timeout refusal rather than killing teardown mid-way. * fix(stow): add opt-in pass horizon for memory decay (#2850) * fix(stow): give memory decay a per-pass horizon so the clock fires The tiered decay clocks were wall-clock only, while admission is per-pass: each /stow admits the findings that pass produced. In a home that stows daily those two rates diverge by the stow cadence, an entry the fleet keeps exercising never reaches 30 days unreinforced, and memory only grows while the pass reports decay evaluated. Give each dated marker an optional unreinforced-pass counter and make both tiers stale at whichever horizon comes first: 10 passes or 30 days for aging, 3 passes or 7 days for perishable. Reinforcement clears the counter and nothing else does, so the existing evidence-based restamp rule stays the only way an entry renews its lease. An absent /N means zero, so entries that stay exercised carry no extra marker bytes, and a rarely stowed home keeps its current behaviour through the unchanged date horizon. * no-mistakes(document): Align stow workflow with dual decay clocks * fix(stow): make the per-pass decay horizon opt-in The unreinforced-pass horizon shipped as a new default archival cadence, which is a product default rather than a restoration of the existing wall-clock contract. Keep the 30-day and 7-day horizons as the only default clock, and put the 10-pass and 3-pass horizons behind an explicit opt-in: config/stow-pass-horizon for the firstmate home, and the file's own header pointer for the public skill. With the opt-in absent no counter is written and no counter is read, so a home that does not ask for it decays exactly as it does today. * no-mistakes(review): Preserve frozen counters and correct archive provenance * test(watcher): stop fixture confirmation budgets racing real child startup (#2876) tests/fm-watcher-lock.test.sh passed in isolation but failed intermittently under full-suite and ambient concurrent load. bin/fm-watch-arm.sh computes its confirmation deadline immediately after forking the real child watcher, so the child's entire fork, exec, lock acquisition and beacon publication has to land inside that wall clock. Two cases shrank that budget to one second, leaving a two-second window for work measured at 3.1-4.9s under CPU oversubscription, so the arm honestly reported "FAILED - no live watcher with a fresh beacon" and their premises collapsed. A third case ran on the production budget, but its child must also execute a registered check before exiting: measured at 1.9-2.3s idle and 9.1-13.1s under load, against an 11s budget. The two cases that must confirm a real child now hold the arm to production's own budget instead of a shrunken fixture one, the immediate-wake case gets an explicit budget with headroom over its measured loaded cost, and the two waits for the arm's typed failure are sized off the largest production default rather than a fixed eight seconds. No bin/ change and no default behavior change: the lock's fail-closed semantics, SIGSTOP handling, stale-heartbeat detection and the arm's typed failures are untouched. Verified 4/4 green at 3x CPU oversubscription (loadavg 75-80) after 3/3 red before the change, and CONTRIBUTING.md records the convention. * fix(bin): deterministically order remote tool paths (#2870) * fix(bin): order discovered tool installs by the shell's own expansion fm_remote_job_compose_operator_path built the asdf and mise install directories with `compgen -G`, which does not sort. Bash sorts glob matches in pathexp.c, on the shell's own pathname-expansion path only; `compgen -G` reaches the same glob_filename through pcomplete.c, which sorts nothing. On bash 3.2 (macOS /bin/bash) and every bash before 5.3 that handed the composition raw readdir order, so which install of a multi-version tool a remote job resolved was decided by directory order on disk rather than by this composition. Expand the globs at the call sites and let the function take the matches, so the composition and the documented portable-PATH contract are the same operation. Quoting the account home at the call site also stops a home whose name contains glob metacharacters from being reinterpreted. The colocated regression pins both the order and the mechanism: bash 5.3 moved sorting into the glob library, so an order-only assertion cannot see the defect there. * no-mistakes(review): Remove source-reading PATH regression guard * fix(bin): prevent routed secondmate work from stranding (#2848) * fix: surface stalled secondmate queues and wake handoffs * no-mistakes(review): Make handoff wakes retryable and stall alerts crash-safe * no-mistakes(review): Prevent duplicate handoff wakes and cover remote delivery * no-mistakes(review): Serialize local handoffs and preserve pre-move wake intent * no-mistakes(review): Serialize teardown with handoffs and retain remote wake confirmation * no-mistakes(review): Reconcile correlated handoff wake delivery after crashes * no-mistakes(review): Keep failed wakes retryable and isolate stall receipts * no-mistakes(review): Reset known-undelivered wake attempts for durable retries * no-mistakes(review): Refuse duplicate sends for unresolved delivery attempts * no-mistakes(review): Atomically restore retryability after reconciled send failures * no-mistakes(review): Serialize delivery confirmation with reconciliation * no-mistakes(document): Document routed wake and stall supervision * no-mistakes(lint): Fix ShellCheck expansion and subshell warnings * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Retire stale wake state and defer pre-move wakes * no-mistakes(review): Secure markers, bind batches, and preserve teardown routes * no-mistakes(review): Preserve unresolved prepared wakes across unrelated handoffs * no-mistakes(review): Preserve prepared wakes before unrelated moving handoffs * no-mistakes(document): Document prepared wake batch ownership * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Make local wake retirement recoverable * no-mistakes(document): Clarify handoff recovery and teardown documentation * fix: make macOS inbox test path portable (#2857) * feat(bin): deliver local steers through durable task inboxes (#2856) * feat(bin): steer local tasks by durable inbox record plus constant doorbell Stage 1 (local steers) of the captain-adopted reframe in data/fm-send-reliability-reframe-s1/report.md: an ordinary fm-send text steer to a task recorded in this home is appended as a sequenced durable record under state/<id>.inbox/ and the terminal receives only one constant self-describing doorbell line, best-effort. The worker acknowledges by moving the record into handled/; the watcher re-rings an unacknowledged message on an idle pane and escalates once as an ordinary stale wake. --resolve-key closes decisions at enqueue time, because the durable enqueue IS delivery to the task's record. bin/fm-task-inbox-lib.sh owns the record format, doorbell line, and re-ring ladder. The typed plane remains for what must reach the terminal itself: lifecycle keys, harness-native slash and codex $-skill invocations, explicit backend targets, and the remote secondmate leg (unchanged until the remote inbox leg ships separately). The composer classifier is demoted from delivery proof to an advisory ring guard that skips only on a proven pending verdict. Verified live against claude, codex, opencode, pi, grok, and muse: each real worker read its record, acted, and acked with the mv (docs/verification/runtime-backends.md "Steering-inbox doorbell"). * docs(verification): flag the grok 1.0.5 composer-matrix staleness observed by the doorbell run * test(captain-hold): read the chat-channel answer from the durable inbox record * test: migrate fm-control's marker contrast to the inbox record and fix macOS wc padding in the tool-update suite * no-mistakes(review): Harden inbox locking, teardown races, and acknowledgements * no-mistakes(review): Serialize watcher actions with inbox acknowledgements * no-mistakes(review): Bound metadata locking and tighten acknowledgement rechecks * no-mistakes(review): Preserve exact inbox bytes and harden delivery recovery * no-mistakes(review): Harden watcher bookkeeping against concurrent inbox teardown * no-mistakes(document): Update inbox and typed-plane documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * revert(pipeline): keep parser-native secondmate marking and the both-failed exit out of stage 1 The CI monitor's fix changed the secondmate marking contract for parser-native invocations (appending the marker after the text) and softened the both-commit-and-marker-failed branch to exit 0. The merge authority ruled the marking question out of scope for this stage-1 transport PR (follow-up: fm-send-secondmate-harness-invocation-r1) and ruled the both-failed case a loud nonzero local failure. Restore both, keeping the monitor's legitimate migrations and hardening. * no-mistakes(document): Document inbox and typed-plane boundaries * no-mistakes(document): Scope backend transport docs to typed plane * no-mistakes(document): Clarify inbox attempt-budget documentation * no-mistakes: apply CI fixes * fix(send): the durable record alone governs the inbox exit status Captain-refined ruling on the F2/Greptile finding: the durable inbox record is what delivers the steer, so pending-reply bookkeeping trouble after a successful enqueue never exits nonzero - a resend-inviting status would make automated callers enqueue the delivered instruction again under a new sequence. With the recovery marker stored the watcher reconciles silently; with the commit and marker both lost the send surfaces a distinct reply-tracking-degraded do-not-resend warning and still exits 0. Nonzero remains only where nothing was delivered (or a decision close needs its manual command). Regression: record durable + both bookkeeping writes lost -> exit 0, one record, no duplicate. * no-mistakes(review): Preserve inbox ordering with drain-all doorbells * no-mistakes(review): Surface unwritable inbox ladder bookkeeping * no-mistakes(review): Silence ladder failures after inbox acknowledgement * no-mistakes(document): Update steering inbox documentation * no-mistakes: apply CI fixes * feat(bin): add fast local lint mode (#2891) * feat: add fast local lint mode * fix: preserve complete fm-lint help * fix: isolate fast lint mode * no-mistakes(document): Clarify lint mode documentation ownership * no-mistakes: apply CI fixes * feat(bin): deliver remote steers through durable inboxes (#2901) * feat(bin): deliver remote secondmate steers through durable task inboxes Stage 2 of the inbox+doorbell steer channel (stage 1: #2856). A remote secondmate steer now crosses fm-on.sh as a durable record written idempotently into the remote home's steering inbox plus a best-effort remote doorbell, and the last typed-payload steer transport is deleted: - fm-remote-secondmate-control.sh cmd_send writes the record via the new fm_task_inbox_write_idempotent and rings the doorbell; it no longer types the payload through an inner fm-send at an explicit pane target. - fm-send.sh routes every remote text steer (harness-native included, which marking already reduced to chat) onto the remote inbox leg, retries the identical leg once on ssh 255, closes --resolve-key decisions at enqueue for remote too, and preserves a marked request's reply expectation when completion stays unknown. The exit-3-as- delivered remap, the 255 do-not-resend trap, and the remote typed submit block are removed. - fm-task-inbox-lib.sh owns the idempotent enqueue: an exact-body re-run lands on the existing record, handled or not, so an ambiguous transport can always be safely re-run. - Tests pin the new contract end to end (record + doorbell + no typed payload across ssh, one-record idempotence under an ambiguous transport, enqueue-time decision close, loud real failures, and the deleted typed-payload behaviors gone), and AGENTS.md plus docs/remote-secondmates.md describe the remote leg's new semantics. * no-mistakes(review): Harden remote inbox delivery against lifecycle races * no-mistakes(review): Enable correlation-preserving remote steer resends * no-mistakes(review): Fail closed on stale correlation resends * no-mistakes(review): Include home context in remote resend commands * no-mistakes(review): Lock and revalidate remote parent routes * no-mistakes(document): Clarify remote steer retry documentation * no-mistakes: apply CI fixes * feat: add persistent Pi supervision branch (#2858) * wip: forked supervision on Pi (checkpoint before docs) * fix(pi-branch): harden mirror delivery, fallback encoding, and session replacement Peek-then-shift mirror flush so a failed append retries instead of dropping; durable mirror cursor commits only after delivery into the branch; the main fallback wake is operational-encoded like every watcher injection; session_shutdown quiesces the generation and session_start re-arms, so /new and /resume no longer kill the branch permanently. Registers the extension in the strict typecheck, adds the dispatch handshake test, the branch extension suite, the bash-level regression suite, the session-start replay test, and the opt-in real-SDK live guard. * test(fixtures): carry the branch-dispatch lib and lease lib into isolated fixtures The watcher extension now imports lib/fm-branch-dispatch.ts and fm-teardown sources fm-lease-lib.sh, so every fixture that copies or symlinks those files in isolation gains the new sibling. * no-mistakes(review): Prevent shutdown wake loss and serialize lease claims * no-mistakes(review): Durably hand off wakes and retain portable leases * no-mistakes(review): Require durable reports and clear disposed branch leases * no-mistakes(review): Enforce per-wake outcomes and quiescent lease cleanup * no-mistakes(review): Require wake acknowledgements and tighten branch lifecycle boundaries * no-mistakes(review): Require complete acknowledgements and replay cleanup failures * no-mistakes(review): Bind supervision to lock ownership and durable delivery * no-mistakes(review): Activate branch lazily after session lock acquisition * no-mistakes(review): Preserve undelivered mirror context across extension rebinds * no-mistakes(review): Acknowledge startup replay only after main delivery * no-mistakes(review): Isolate replay metadata from untrusted digest content * no-mistakes(review): Reject duplicate reports for active wake sequences * no-mistakes(review): Retain failed fallbacks and deduplicate outcome replay * no-mistakes(review): Deduplicate durable outcomes and cache delivery receipts * no-mistakes(review): Anchor wake sequence matching to outcome fields * no-mistakes(document): Clarify Pi supervision durability contracts * no-mistakes(lint): Fix ShellCheck issues in branch supervision scripts * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * refactor(pi-branch): collapse to confused-agent-grade guards per captain decision Captain decision A: the lease/actor guards target the CONFUSED-AGENT threat model bin/fm-gate-refuse-lib.sh already documents; adversarial-grade separation is impossible in the shared-process design and is filed as separate follow-up work. Rip out the machinery that chased it: the generation fence and shell-provenance markers, the wrapper-tagged ancestry walks, guard auto-claim with per-script release traps, the pending-wake files and ack-receipt correlation (the durable wake queue already re-presents anything unacknowledged), the delivery-receipt store with contiguous cursor advancement, the session-start replay-metadata channel, and the branch tool quiescence counters. Keep the behaviors the board requires, each on its simplest implementation: lazy per-action session-lock ownership (cold start activates after the lock lands; a secondary session stays inert), mirror durability across extension rebinds via the durable cursor, replay-exactly-once from the one read cursor, the awaited operational-encoded fallback, per-generation stray-lease cleanup, session-lock-bound lease liveness (a recycled pid or a non-Pi home never honors a leftover lease), the loud accidental-override guards (readonly actor prelude, cross-actor claim refusal), and the role-partition refinements (no forced teardown, no direct relaunch for the branch). Default-on-for-Pi is unchanged. * no-mistakes(review): Enforce lock ownership and serialize lease mutations * no-mistakes(review): Synchronize guard cleanup and bind leases to lock owner * no-mistakes(review): Report outcomes before acknowledging durable wakes * no-mistakes(review): Restrict leases to Pi and instruct main claims * no-mistakes(review): Reject malformed lease locks and torn outcome tails * no-mistakes(review): Validate complete outcome tails before appending * no-mistakes(review): Guard branch side effects across session replacements * no-mistakes(document): Update Pi supervision durability and lease documentation * no-mistakes(lint): Suppress intentional nested-shell expansion warning * no-mistakes: apply CI fixes * fix(pi-branch): authorize lease releases by caller * fix(lint): break redundant source-analysis path in fm-lease-lib.sh fm-lease-lib.sh's lazy fallback source of fm-wake-lib.sh gave ShellCheck's --external-sources traversal a second path into an already 1540-line file that fm-send.sh and fm-teardown.sh also source directly, blowing up the recursive analysis past CI's lint timeout. Mark it a source=/dev/null analysis boundary, matching the existing fm-task-inbox-lib.sh convention. Also restores bin/fm-lint.sh and tests/fm-lint.test.sh to the shared serial-lint definition (dropping an unrelated parallel-sharding change that was itself hanging and masked this root cause). * no-mistakes(document): Correct lease caller-authorization documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(bin): parallelize startup network sweeps (#2927) * feat(bin): parallelize session-start remote secondmate network sweeps Run per-secondmate liveness and convergence probes concurrently and overlap clone refresh, while replaying each mate's fail-closed diagnostic in original order. Ignore scratchpad* so untracked scratch no longer blocks remote sync. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(document): Document parallel startup network sweeps * no-mistakes(lint): Fix empty environment assignment lint warning * no-mistakes: apply CI fixes --------- Co-authored-by: Cursor <cursoragent@cursor.com> * test: handle absent watcher wake queues (#2845) * fix(tests): count declared-pause wakes without crashing on an absent queue The exited-declared-pause case counts queued stale wakes by handing state/.wake-queue straight to awk. A watcher that queues nothing never creates that file, and awk aborts on a missing path before its END rule runs, so the count collapses to the empty string. The next comparison then fails as an integer-expression error and surfaces as a wake flood with no number, hiding the real contract breach the following grep names. Read the queue the way the drain-count assertion at the end of this file already does: silence awk's open error and default an absent queue to zero. Applied to all four counts in this case, including the live external-decision gate pair whose queue an acknowledged drain can also leave behind. An absent queue now reports "did not use the bounded paused recheck", while a genuine flood still fails with its real count. Fixes #2628 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * style(pi): distinguish routine and captain supervision merge notes by icon (#2934) * style(pi): restyle supervision merge notes with a sailboat and matching pad Secondary-session notes were flush against the TUI edge and fully tinted. Use the sailboat prefix, Pi's default outputPad, boat-only color, and dim remainder so they sit like real messages. * style(pi): distinguish routine and captain merge notes by icon only Visible notes now lead with a sailboat or anchor, then only the dim outcome. Drop the branch-merged wording and verdict brackets so the icon is the only kind signal. * docs(pi): add the approved multi-brain architecture poster (#2938) The markdown contract stays the owner; the still is only the visual of the idea. * feat(pi): default branch supervision and route heartbeats (#2939) * fix(bin): bound remote job worker supervisor restarts (#2942) * fix(bin): bound remote worker supervisors * no-mistakes(review): release incumbent supervisor before starting its replacement * no-mistakes(review): wait out a healthy same-root supervisor instead of replacing it * no-mistakes(review): narrow remote worker change to restart accounting only * no-mistakes(document): clarify supervisor restart guard is a lifetime total * fix: safely split supervision wake handling by actor (#2953) * feat(bin,pi): per-actor wake consume, silent success gating, merge-poll dedup Three related fixes to the shared wake-drain and Pi supervision-branch dispatch machinery so a routine success is never main-blocking and a mixed queue can safely split between actors. 1. Successful routine results no longer create main-blocking wake rows. fm-startup-network.sh only enqueues a check: startup-network wake when the deferred result is actionable (state is not "done", or the report carries a bootstrap-diagnostics actionable prefix); a clean success stays durable in the report file without ever waking the agent. 2. Per-actor wake-drain consume contract. bin/fm-wake-drain.sh now scopes presentation and --ack-through to the current actor (bin/fm-lease-lib.sh's fm_lease_actor): main keeps the original whole-queue cutoff behavior, unaffected. A branch actor (FM_SUPERVISION_ACTOR=branch, set only inside the Pi supervision branch's own bash tool calls) is scoped to an explicit eligible-row snapshot instead of a cutoff comparison, so it can never remove a row it was not granted - the fix for the swallow risk that used to force an all-or-nothing whole-queue fallback to main. .pi/extensions/lib/fm-branch-dispatch.ts's scopeForUnreadWake is the single owner of eligibility: a check-kind row (merge-confirmation polls, Relay mentions, credential/auth failures) is now excluded rather than vetoing the whole scan for a non-heartbeat wake, while a heartbeat review keeps its original all-or-nothing rule unchanged. writeEligibleRowsSnapshot publishes the exact eligible sequence numbers before every branch prompt; fm-primary-pi-watch.ts's offer still refuses a check-kind trigger outright so a main-only close is never itself routed to the branch. 3. A repeat identical merged-PR-poll result for an already-notified task is absorbed instead of enqueued again. A poll's own retirement state is scoped to one registration and cannot see a prior registration's outcome, so a task re-registered after its merge was already surfaced would otherwise wake main a second time for the same event. bin/fm-pr-lib.sh's new per-task pr-poll-merge-notified marker survives across re-registrations to catch that case; the first notification for a task still reaches main unchanged. Regression tests colocated in tests/fm-startup-network.test.sh, tests/fm-wake-queue.test.sh (including the mixed-queue no-swallow property), tests/fm-pi-branch-extension.test.sh, and tests/fm-pr-check-security.test.sh. docs/watcher-continuity.md and docs/pi-supervision-branch.md updated for the new contracts. * no-mistakes(review): Bind merge deduplication to canonical PR identity * no-mistakes(review): Serialize wake row ownership across main and branch * no-mistakes(review): Bind branch grants and deduplicate within actor claims * no-mistakes(review): Fallback main-owned wake claims to main delivery * no-mistakes(review): Clarify silent startup success guidance * no-mistakes(review): Release residual branch grants after settled prompts * no-mistakes(review): Reject truncated wake rows as corrupted * no-mistakes(document): Document per-actor routing and silent startup success * no-mistakes(lint): Fix ShellCheck findings in wake grant and startup test * no-mistakes: apply CI fixes * fix(pi): hide branch outcomes tool rows in Calm (#3024) * Hide branch outcome tool in Pi Calm * no-mistakes(review): Preserve stock outcomes rendering and document tool audit * no-mistakes(review): Document branch read tool audit disposition * no-mistakes(review): Match stock outcomes output sanitization * no-mistakes(document): Document Calm custom-tool visibility * fix: bind no-mistakes attestations to PR head (#3027) * fix: delegate no-mistakes PR gate to pinned action * no-mistakes(document): Document commit-bound no-mistakes attestations * feat(pi): add persistent supervision branch model selection (#3028) * feat(pi): let operators pin a cheaper supervision-branch model Supervision is an easier job than the captain's own conversation, so the Pi supervision branch does not need main's model. A new /supervision-model command opens Pi's own selector over Pi's own catalog of credentialed models, plus a "Follow main" entry, and persists the pick as one <provider>/<model-id> line in this home's gitignored config/supervision-branch-model. Firstmate keeps no model catalog of its own. The branch resolves the pin at every branch build - the first wake of a cold start and the reopen after /new, /resume, /fork, or reload - so the choice survives all of them, and picking also releases the live branch so the next wake reopens the same persistent branch conversation under the new model. An absent, unreadable, or unparseable file means no pin and keeps today's behavior byte for byte: no model option is passed and Pi picks the branch's model exactly as before. A pin naming a model Pi cannot hand back is never silently downgraded onto main's model: the branch refuses to build and the wake falls back to the captain-facing main path naming the unusable pin, which is the extension's existing failure direction. The choice is home-local and not part of secondmate inherited configuration, matching the Pi Calm preference precedent. docs/configuration.md owns the operator-facing schema. Portable regressions cover pin-present on create and reopen, pin-absent default, the command's persistence, cancellation, and live rebind, and both unusable and unparseable pins. The opt-in real-SDK guard proves the vendor surface the pin reads and that an explicit model wins over the model a reopened session recorded. * no-mistakes(review): Fix supervision model runtime and rebind races * no-mistakes(review): Restrict supervision picker to isolated runtime models * no-mistakes(document): Document supervision branch model selection * fix(pi): make the supervision model pin authoritative on every reopen Clearing the pin with "Follow main" removed the file but the next branch build reopened the persistent branch session with no explicit model override, so Pi restored the model that session had recorded - the old pinned model - while the command reported that the branch now follows main. The same gap meant an absent pin did not reliably mean same-model-as-main once a home had pinned once. The pin file's current state now decides the model on every branch build, create and reopen alike, overriding Pi's session-state restore. With a pin, that model. With no pin, main's own current model is applied explicitly, tracked from the contexts Pi already hands the extension plus its model_select event, since the branch is built at wake time with no context of its own. Only when main's model is unknown, or this home's stored credentials cannot run it in the isolated branch runtime, does a build fall back to passing no override at all, which is the behavior from before the pin existed; the branch is never refused over model choice. The command's notification now reports the model actually applied, and says plainly when clearing the pin could not apply main's model instead of claiming a change that did not take effect. No credential handling changes: the branch still relies entirely on the stored credentials its own runtime already holds, and the picker stays restricted to models that runtime can resolve. Colocated regressions cover pin present on create and reopen, clearing the pin returning a reopened branch to main's model and specifically not the old pinned one, an unparseable pin behaving as no pin, and the unknown-main-model fallback to no override. * no-mistakes(review): Make unpinned supervision follow main model changes * no-mistakes(document): Correct supervision model documentation * feat(pi): let /supervision-model pick branch reasoning effort (#3079) * feat(pi): let /supervision-model pick the branch's reasoning effort Supervision is an easier job than the captain's own conversation, so the Pi supervision branch does not need main's reasoning effort any more than it needs main's model. /supervision-model now settles both in one flow: the existing model picker, then a follow-up effort picker built from Pi's own supported thinking levels for the model just chosen. Firstmate keeps no effort catalog of its own; the menu, the clamp, and the vocabulary all come from Pi. The pick persists as one line in this home's gitignored config/supervision-branch-effort, independent of the model pin: a captain may pin a model, an effort, both, or neither. The effort pin's current state decides the branch effort on every branch build - the first wake of a cold start and the reopen after /new, /resume, /fork, or reload - and overrides Pi's restore of whatever level a reopened branch session recorded, which is what keeps "Follow main" honest. With no pin, main's own current effort is applied explicitly and followed live through Pi's thinking_level_select event, the same way an unpinned branch already follows main's model, and the two selections now share one build revision so either change invalidates an in-flight build. The branch is never refused over effort. Pi owns the clamp, so a pinned level the branch's model cannot run becomes that model's nearest supported level while the captain's raw pick is kept for a model that supports it, and the command reports the level the branch will really run at rather than the raw pin. A token Pi would not recognize at all is treated as no pin rather than passed to that clamp, which would otherwise collapse a typo into the model's lowest level. Only when main's effort cannot be read either does a build pass no effort override at all, which is the behavior from before this file existed. Pi's own effort vocabulary is pinned by a bidirectional type assertion against Pi's getThinkingLevel return type, so the tracked strict typecheck against the installed package fails the moment Pi adds or removes a level. docs/configuration.md owns the operator-facing schema for both pins. Portable regressions cover the pin on create and reopen, model-only and effort-only pins working independently, clearing a pin returning the branch to main's effort, live-follow of a mid-session change, the clamp, an unrecognized token, the unknown-main-effort fallback, and the command's two-step flow, persistence, cancellation, and honest reporting. The opt-in real-SDK guard proves the vendor surface all of that rests on, and also repairs a pre-existing gap that left it unable to load the extension at all. * no-mistakes(review): Resolve effective branch effort honestly * no-mistakes(document): Clarify Pi-owned effort picker behavior * fix: keep routine supervision noise out of captain chat (#3093) * fix(supervision): silence empty board closes and decouple the heartbeat Two unrelated sources of noise put routine supervision events in the captain's chat. An empty Lavish board close - the captain reads a review surface, says nothing, and closes it - became a check wake whose entire content was that nothing happened. Suppress it at its source instead of routing it anywhere: the generic runner gains a `silent` adapter seam mirroring the existing `terminal` one, and the Lavish adapter answers it for exactly one positively-determined shape, an `ended` session carrying no queued content block. A silenced result is recorded durably handled so it does not return on a later reconcile. Everything else announces unchanged - a `Send & End` close carrying the captain's real answer, an `ended` result still carrying content, a waiting or missing session, an unreadable result, and every adapter that implements no `silent` command at all. The keyed-answer feed is untouched, so suppressing an announcement never suppresses the captain's own answer. A fleet heartbeat was deferred to main merely because some unrelated check row happened to be sitting unread, which put a routine fleet review in the chat for a reason that had nothing to do with the fleet. A check row is permanently main-owned, so it is now excluded from a heartbeat claim rather than vetoing the scan, exactly as in every other mode. What all-or-nothing guarantees is unchanged: the branch takes every branch-ownable unread row or none of them, and an unresolvable task-local row, an unknown row kind, or an unreadable queue still defers the whole review to main. Main is still woken for the check on that check's own triggering close, so nothing starves. Main-only classes are unchanged and now each covered by a test: Relay mentions, credential failures, merge confirmations, real board answers, and watcher-failure repair. The per-actor acknowledgement and no-cross-swallow properties are untouched. * no-mistakes(review): Fail closed on all Lavish content headers * no-mistakes(review): Suppress false unacknowledged status for silenced results * fix(bin): stop a correlation token from hiding and stranding decisions (#1967) * fix(classify): read the decision fold through a correlation token status_line_verb stripped a trailing [key=...] from a status line's prefix but left everything else glued to the verb, so a line carrying the correlation token bin/fm-pending-reply-lib.sh embeds and a secondmate echoes back matched no arm of _fm_decision_fold_line. Such a line folded as ordinary status in both directions: a needs-decision or blocked opener never opened its key, and a resolved or captain-held closer never closed one. The same glued verb also hid correlated done and blocked lines from status_is_captain_relevant and status_is_terminal_verb, and let correlated working and resolved lines leak through the free-text fallback the nonterminal guard was meant to stop. The verb parse now walks whole words and drops only a token of the exact shape a firstmate library writes - corr=<16 hex>, plus the bracketed form bin/fm-secondmate-report.sh emits - before or after the key token, unkeyed, or doubled. An arbitrary name=value word is deliberately NOT skipped: skipping unknown tokens would let free text carrying an equals sign reduce to a bare verb and impersonate a transition, which is the takeover the strict parse and _fm_decision_key_transition_allowed exist to prevent. A prefix with no corr= substring is returned byte-for-byte as before, so every line without a token keeps its exact historical verb. FM_OPEN_DECISIONS_FOLD_VERSION goes to 3, because every cursor persisted under the previous reading carries an open set computed while correlated lines were invisible and must be rebuilt from byte 0. Measured over a real 383-line status log: 254 lines keep byte-identical captain-relevance, pause, terminal-verb and captain-held verdicts, and all 129 changed lines carry a valid token - 14 correlated done/blocked/ needs-decision lines become captain-visible, and 20 correlated working/resolved lines stop being escalated on prose alone. * fix(review): Captain, block token-first decision impersonation * fix(document): Clarify normalized status verb ownership * fix(classify): reconcile the correlation-token read with the tag-stop parser Rebasing onto main put this change beside #2280, which made verb parsing stop at the first "[name=value]" tag. Both edit status_line_verb with different intents, so the resolution keeps both rules rather than letting one overwrite the other: - #2280's tag stop is kept verbatim and now owns every BRACKETED tag, including the "[corr=...]" form fm-secondmate-report.sh writes. The bracket-unwrapping arm this branch had added to the token test is therefore removed as unreachable. - This branch's token walk is kept and narrowed to the UNBRACKETED token fm-pending-reply-lib.sh writes, which the tag stop does not reach. Two consequences of standing beside #2280 rather than before it: The fold version had collided at 4: #2280 spent it on the tag-stop parser and this branch had spent it on the token read. A cursor persisted under #2280's reading predates this one and must still be rebuilt, so the version moves to 5. A bracketed impostor is dropped from the malformed-token list. On main today "resolved [corr=deadbeef] [key=victim]:" already reads as the bare verb, as does "resolved [anything at all] [key=victim]:", because the tag stop ends the parse at the opening bracket regardless of content. That is #2280's reviewed contract; asserting otherwise here would narrow it. The unbracketed impostors it owns stay strict and still fold as prose. Adds a consumer test for the two verb-string case arms that postdate this branch: fm-supervise-daemon.sh's transient-stale arm and fm-crew-state.sh's map_log_state. * fix(review): Captain: Seed cursor migration fixture with version four * fix(document): Clarify voice status normalization ownership * fix(bin): Cursor-Park unter Pi-Host ohne Cursor-Identität stilllegen (#3115) * fix(bin): Cursor-Park unter Pi-Host stilllegen. pi-cursor-sdk lädt .cursor/hooks.json in die Pi-Sitzung und parkte einen zweiten Watcher; das erzeugte rearm-resurface und brach laufende Rückfragen ab. Bei PI_CODING_AGENT=true beendet der Park sofort, native cursor-agent Primaries bleiben unverändert. * fix(bin): Cursor-Park trotz PI-Leak nur ohne Cursor-Identität stilllegen. Stand-down gilt nur bei PI_CODING_AGENT=true ohne CURSOR_AGENT und ohne CURSOR_INVOKED_AS. Handgestartete cursor-agent Primaries mit geerbtem PI-Marker parken weiter. * no-mistakes(document): Document Cursor park Pi-host stand-down * fix(bin): no-mistakes-Mindestversion auf 1.46.0 anheben. Die PR-Attestierung verlangt ab 1.46.0 strukturierte Pipeline-Schritte; der Bootstrap-Floor blieb bei 1.31.2 und ließ zu alte Builds zu. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> * fix(pi): make supervision model picker searchable and scrollable (#3099) * fix(pi): make /supervision-model's model list bounded and searchable Pi's generic extension selector renders every option at once with no search box, so a real eligible catalog ran off the top of the terminal. The model step now draws the same rows through Pi's own SelectList - the bounded scrolling primitive behind Pi's /model picker - with Pi's own Input and fuzzy filter above it for search, keeping 'Follow main' first, the branch-runtime eligibility filter intact, and the pick branch-only. Pi's ModelSelectorComponent is deliberately not reused: its selection handler writes the captain's default model through Pi's settings manager, which would move main's conversation as a side effect of pinning the branch. The effort step's menu is a handful of levels and stays on Pi's plain selector dialog. * no-mistakes(document): Clarify supervision picker documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(docum…
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Intent
Give firstmate's guarded merge path a GitLab route, so a GitLab merge request lands through the same recorded script and the same refusals a GitHub pull request already gets, with every pre-merge condition (open state, mergeable status, no conflicts, resolved blocking discussions, and a successful pipeline at the exact live head) enforced by the script rather than checked by hand, the instance taken from the merge request URL rather than hardcoded, and the existing GitHub behaviour left byte-identical.
What Changed
bin/fm-pr-merge.shno longer refuses a GitLab merge request URL: it now dispatches on the parsed provider, sending a merge request toglab mr merge <n> -R https://<host>/<path> --sha <verified-head> --yeswith the project URL andGITLAB_HOSTrebuilt from the parsed host and path so no instance is hardcoded, and adding no merge-method flag so the project's own merge method applies. The GitHub branch still runsgh-axi pr merge <n> --repo <owner>/<repo>with the--squashdefault and explicit merge-method flags preserved, and is byte-identical to the base commit in every replayed invocation except the one authorised narrowing described below.glab mr view -F jsonread parsed withjqbefore merging: open state,detailed_merge_statusmergeable,has_conflictsfalse,blocking_discussions_resolvedtrue, and a successful head pipeline at the exact live head. Every failing condition is reported rather than just the first, an unreadable or malformed payload refuses instead of merging, a recordedpr_head=that disagrees with the live head is reported as stale and the live head is what gets verified and bound to the merge, and a missingglaborjqrefuses before any state is recorded or poll armed.-R(for example-yR) in addition to--repo/-R, and rejects--shaon GitLab so the head comes only from the live read.tests/fm-pr-merge.test.shadds glab/jq fixtures covering routing, host derivation, per-condition and combined refusals, stale head reporting, tool absence, and unchanged GitHub arg handling;tests/fm-pr-check-security.test.shupdates its GitLab case to assert the merge path reads through glab instead of erroring. Docs (README.md,docs/architecture.md,docs/scripts.md) and thedocs/gitlab-merge-watch.mdverification record were updated for the new merge route.tests/fm-remote-job.test.shsegfaulted its own worker in CI. Its shutdown case signalled in an unpaced loop, delivering signals fast enough to corrupt the signalled shell's pending-trap bookkeeping and kill it part way through the shutdown, which reported a shutdown defect the worker does not have. The burst is now bounded and the follow-up signals paced, so the case still pins the same property: ten back-to-back signals all land inside the shutdown's first file operation, and restoring the default disposition still leaves the ownership lock behind every run. Test-only, no product behaviour changed.Authorised GitHub narrowing
This change NARROWS GitHub behaviour. That was authorised deliberately after being raised during validation, rather than slipping in by accident.
What it now refuses is extra merge arguments containing a bundled short-option cluster that includes
R, for example-dR other/repo. The forge CLIs expand such a cluster one character at a time, so it carries--repo other/repo, and that later value wins over the repository the URL named. Before this change,fm-pr-merge.sh <task> <github-url> -- -dR other/reporeachedgh-axi pr merge 12 --repo example/repo --squash -dR other/repoand exited 0 withpr=recorded and the merge poll armed. It now exits 1 witherror: extra merge arguments must not override the repository, records nothing, and invokes no forge merge command. Every other replayed GitHub invocation is byte-identical to the base commit.Closing that hole honours the existing rule rather than departing from it. The file header already forbids
--repoand-Rbecause the repository must come only from the URL, so a bundled cluster carrying a repository override was never legitimate behaviour to preserve: it was that guard being evaded. Redirecting a merge to a repository the URL does not name is exactly what the guard exists to prevent.The guard covers both providers rather than GitLab alone, because a script that guards one provider and not the other is a trap for whoever reads it next, who would reasonably assume symmetry. On GitLab the same evasion is worse: it would redirect the instance, not just the project.
The refusal is pinned on both paths by
test_bundled_repo_override_args_refuse_before_recordingintests/fm-pr-merge.test.sh, which asserts exit 1, the refusal wording, nopr=recorded, no armed merge poll, and no forge merge command invoked, with a control case proving a cluster carrying no repository override still reaches the forge. Both assertions were confirmed to have teeth by narrowing the guard back to a bare-Rand watching each path fail.Risk Assessment
✅ Low: Every intent criterion is met by fail-closed source logic at a single forge-invocation site with behavioral test coverage that fails before the fix, the three round-1 findings are genuinely resolved and independently verified (guard completeness checked against pflag cluster expansion, both tool-absence cases now host-independent), and the only residual items are one user-authorized deviation from "GitHub byte-identical" and a pre-existing self-fulfilling test assertion, neither of which blocks merge.
Testing
I ran the whole tests/fm-pr-merge.test.sh file (24 cases, all pass) plus the two GitLab cases from tests/fm-pr-check-security.test.sh in isolation, then went past unit tests to drive the real bin/fm-pr-merge.sh as an operator would and captured a CLI transcript showing, for 18 GitLab scenarios, the exact forge command line invoked, the refusal text, the recorded task metadata, and whether the merge poll was armed: a green merge request reaches glab mr merge with GITLAB_HOST and -R taken from the URL and --sha bound to the live head after pr= is recorded and the poll armed, each of the six pre-merge conditions refuses independently and names itself, all six report together, a stale recorded pr_head is reported and the live head verified, and glab or jq absent refuses before anything is recorded. For the byte-identical GitHub constraint I replayed 18 GitHub invocations against a git archive of base commit 52d20f1 and against this branch and the transcripts diffed to nothing; the bundled -R cluster set is the sole difference and matches the fix commit's wording exactly. The refusal is already pinned on both providers by the existing test, so I added no duplicate and instead proved both assertions have teeth by narrowing the guard back to a bare -R in a scratch copy, where each path fails. No UI surface is involved - this is a shell CLI change, so CLI transcripts are the end-user-visible artifact. Overall result: green, no actionable findings; the pre-existing timing-sensitive watcher case was left untouched and not run, as instructed.
Evidence: GitLab merge route CLI transcript (18 scenarios, real bin/fm-pr-merge.sh)
GitLab: a merge request that satisfies every pre-merge condition ==== merge request state on the instance: open, mergeable, no conflicts, discussions resolved, head aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa, pipeline success at aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa $ fm-pr-merge.sh task-x1 https://gitlab.example/group/subgroup/project/-/merge_requests/7 exit status: 0 stdout: | armed: state/task-x1.check.sh | merged! stderr: | verified: https://gitlab.example/group/subgroup/project/-/merge_requests/7 is open and mergeable, with a successful pipeline at head aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa forge commands invoked (GitLab): $ GITLAB_HOST=gitlab.example glab mr view 7 -R https://gitlab.example/group/subgroup/project -F json $ GITLAB_HOST=gitlab.example glab mr merge 7 -R https://gitlab.example/group/subgroup/project --sha aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa --yes forge commands invoked (GitHub): none recorded in task meta: pr=https://gitlab.example/group/subgroup/project/-/merge_requests/7 merge poll armed: yes [...refusals for state / detailed_merge_status / has_conflicts / blocking_discussions_resolved / pipeline status / pipeline sha, the all-conditions-fail report, the stale recorded head notice, the unreadable state refusal, glab-absent and jq-absent refusals, the --sha refusal, extra flag forwarding, and the propagated merge failure follow in the file...]Evidence: Authorised GitHub narrowing: base 52d20f1 vs this branch on the bundled -R cluster
$ fm-pr-merge.sh task-x1 https://github.com/right/repo/pull/6 -- -dR wrong/repo -exit status: 0 -stdout: - | armed: state/task-x1.check.sh +exit status: 1 stderr: + | error: extra merge arguments must not override the repository forge commands invoked (GitLab): none -forge commands invoked (GitHub): - $ gh-axi pr merge 6 --repo right/repo --squash -dR wrong/repo -recorded in task meta: pr=https://github.com/right/repo/pull/6 -merge poll armed: yes +forge commands invoked (GitHub): none +recorded in task meta: nothing (no pr= line) +merge poll armed: no (the GitLab path moves from exit 2 "invalid PR merge request" to exit 1 "extra merge arguments must not override the repository"; the control case, a short cluster carrying no repository override, is unchanged and still reaches the forge)Evidence: Mutation check: the bundled -R guard's coverage has teeth on both providers
guard narrowed from -R to -R in a scratch copy; the guard line is the only difference between the trees --- tree: unmutated | runner: GitHub sub-case first --- ok - fm-pr-merge refuses a bundled short-option repo override and forwards other short flags exit status: 0 --- tree: unmutated | runner: GitLab sub-case first --- ok - fm-pr-merge refuses a bundled short-option repo override and forwards other short flags exit status: 0 --- tree: guard narrowed to bare -R | runner: GitHub sub-case first --- not ok - bundled-repo-override: fm-pr-merge should refuse a bundled repo override: expected exit 1, got 0 exit status: 1 --- tree: guard narrowed to bare -R | runner: GitLab sub-case first --- not ok - bundled-repo-override-gitlab: fm-pr-merge should refuse a bundled instance override: expected exit 1, got 0 exit status: 1Evidence: GitHub parity inputs: 18 invocations on this branch and on base 52d20f1 (diff is empty)
$ diff -u github-base.norm github-head.norm IDENTICAL (18 GitHub scenarios: default squash, --merge, --method=merge, --rebase, extra args, caller --sha abc123, -d, --repo, --repo=, bare -R, deep org path, malformed URL, traversal, bad scheme, bad task id, missing meta, merge failure, no args. Base tree exported withgit archive 52d20f1; sandbox paths normalised before diffing. Companion file: github-base.txt)Evidence: tests/fm-pr-merge.test.sh (24 cases)
Evidence: The two GitLab cases from tests/fm-pr-check-security.test.sh, run in isolation
$ bash subset-check-security.test.sh # test_gitlab_merge_watch + test_gitlab_merged_poll_retires ok - GitLab merge requests are followed on any instance and never wake falsely ok - GitHub and GitLab exact merged results share one retirement path exit: 0/tmp/no-mistakes-evidence/01M0KJMZ2XCNBM4E4Q8W34X93K/drive-merge.sh) - Outcome: 🔧 2 issues found → auto-fixed ✅ across 2 runs (1h5m47s)Pipeline
Updates from git push no-mistakes
✅ **intent** - passed
✅ No issues found.
✅ **Rebase** - passed
✅ No issues found.
tests/fm-pr-merge.test.sh:679-mirror_path_withoutbuilds the tool-free PATH only fromBASE_PATH(the developer's real PATH, captured at line 34) and never includes the case'sfakebin, so thetool=jqiteration runs with noglabmock visible. It therefore depends on a realglabbinary being installed on the host. On a machine without glab,bin/fm-pr-merge.sh:109-118finds both tools missing and printserror: merging a GitLab merge request requires glab and jq on PATH, soassert_grep "...requires jq on PATH"(line 688) never matches and the test fails with a misleading message. Note the sibling helper intests/fm-pr-check-security.test.sh:2843-2854deliberately mirrors$dir/fakebinalongsideBASE_PATHfor exactly this reason. Fix: symlink$case_dir/fakebin/glabinto$case_dir/nojq/(or mirror the fakebin too, minus the omitted tool).tests/fm-pr-check-security.test.sh:2901- The new merge assertion intest_gitlab_merge_watchrunsrun_merge_entrywithPATH="$dir/fakebin:$BASE_PATH"whereBASE_PATHis the deliberately restricted/usr/bin:/bin:/usr/sbin:/sbin(line 23), andmake_case's fakebin (lines 65-102) providesgh,gh-axi, andglabbut nojq. Where jq is not installed under those four directories (e.g. Homebrew/opt/homebrew/bin/jqon Darwin, which this repo supports perfm_pr_file_mode),fm-pr-merge.shexits at the tool precondition with an emptyglab.log, sogrep -qF "mr view 7 -R https://gitlab.example/..."fails with "merge wrapper did not read the merge request through glab at its own instance" - a failure that names the wrong cause and never exercises the intended unparseable-JSON refusal.tests/fm-pr-merge.test.sh:45guards this with an explicitJQ_BINprecondition; this file has none. Fix: symlink the real jq into$dir/fakebin(asadd_glab_mockdoes withln -sf "$JQ_BIN"), or add the same fail-fast precondition.bin/fm-pr-merge.sh:72-reject_repo_overridesonly matches--repo,--repo=*,-R, and-R?*, so a bundled short-option cluster slips through:fm-pr-merge.sh <id> <mr-url> -- -yR https://other.example/g/ppasses the guard, and pflag (which both glab and gh use) expands-yRto--yes --repo https://other.example/g/p, with the later-Rwinning over the script's own-R "$PROJECT_URL". That defeats the invariant the file header states as absolute ("Extra args must not include --repo or -R because the repository comes only from the URL"), and this change makes the guard newly load-bearing across instances, since a GitLab-Rcan name any host rather than only another repo on one forge. Practical impact on the GitLab arm is limited because--sha "$FM_PR_MERGE_HEAD"would then have to match the redirected project's MR head, and the same cluster form (-sR owner/repofor gh) is pre-existing and unguarded on the GitHub arm, where no--shabinding exists. Fix in the shared helper: also refuse any single-dash argument whose short-option cluster containsR.bin/fm-pr-merge.sh:221- Requiringhead_pipeline.status = successANDhead_pipeline.sha == .shais exactly the condition the intent marks required, and the author documents the main consequence honestly in docs/gitlab-merge-watch.md:257-258 ("A project that runs no pipeline at all therefore cannot merge through this path"). Two consequences worth being aware of, neither of which is a defect and both of which fail closed rather than merging unverified: (1) a GitLab project with no CI can never merge through this path, while the GitHub arm still merges no-CI repos - a deliberate asymmetry with the "same refusals as GitHub" framing; (2) on projects using merged-results pipelines,head_pipeline.shais the merge-ref commit rather than the MR's source head, so the sha-equality check would refuse even a green pipeline. Detached merge-request pipelines and plain branch pipelines are unaffected. No action needed for intent conformance; flagging so the limitation is a known one rather than a surprise refusal.🔧 Fix: reject bundled -R clusters and make tool-absence cases host-independent
2 infos still open:
bin/fm-pr-merge.sh:83- The sharedreject_repo_overridestightening means GitHub behavior is no longer literally byte-identical:fm-pr-merge.sh <id> <github-url> -- -dR wrong/repopreviously reachedgh-axi pr merge <n> --repo right/repo --squash -dR wrong/repo(where pflag's later-Rwon) and now refuses with exit 1. The intent says "the existing GitHub behaviour left byte-identical", so noting the deviation for traceability only: the user directed exactly this in the round-1 fix instructions ("the only intended change there is that a previously accepted malformed invocation now refuses; do not otherwise alter GitHub behaviour"), and I verified it is a strict tightening. Running the guard's case arms over 27 forms shows only single-dash arguments containingRnewly refuse;--repo/--repo=*,-R,-R?*refuse exactly as before, and-d,-ds,-y,-m,-F,-A,-,--sha,--squash,--method=merge,--remove-source-branchare all still forwarded. No existing GitHub or GitLab assertion in either suite changes outcome. No action needed.tests/fm-pr-merge.test.sh:226-run_pr_mergespecial-cases theunsafe-url-segmentcase: when the script exits 2 it printserror: PR URL must match https://github.com/<owner>/<repo>/pull/<number>to its own stderr (captured into$case_dir/stderr) and returns 1. That string exists nowhere inbin/, sotest_rejects_unsafe_url_segments_before_recording's two assertions at lines 354-355 (expect_code 1and the message grep) only re-read what the harness itself just emitted. The case's other assertions (nopr=recorded, no poll armed, nopr merge) are real and still catch an acceptance regression, so this is a pre-existing shim from an earlier change rather than something this diff introduced. It is now stale, though: the fabricated message asserts a GitHub-only URL contract that this change replaces with a two-forge one, and the changedtest_malformed_url_refuses_before_merge(lines 328-330) already asserts the genuine contract, exit 2 witherror: invalid PR merge request. Dropping the shim and switching this case toexpect_code 2plus the real message would remove the self-fulfilling assertion without losing coverage.🔧 **Test** - 2 issues found → auto-fixed ✅
tests/fm-pr-check-security.test.sh:2512- tests/fm-pr-check-security.test.sh casetest_custom_snapshot_cleanup_on_signalfails on this host: it waits at most 100 × 0.02s = 2s for the watcher to spawn the custom check child, and this host needs longer (the whole suite runs 906s here versus the 250417ms recorded in bin/fm-test-run.sh, ~3.6× slower). It is pre-existing and unrelated to this change - it fails 3/3 on a checkout of the base commit 52d20f1 and 3/3 on this branch, and passes when the ceiling is raised to 60s. Left untouched to avoid widening this commit's scope; the author may want to decide whether that bounded wait should scale with host speed.bin/fm-pr-merge.sh:76- The only GitHub-visible behaviour difference from the base commit, out of 21 invocations replayed against both trees:fm-pr-merge.sh <task> <github-url> -- -dR other/repopreviously reachedgh-axi pr merge 12 --repo example/repo --squash -dR other/repo(exit 0, pr= recorded, poll armed), and now refuses with "extra merge arguments must not override the repository" (exit 1, nothing recorded). It is deliberate (the review commit's bundled-cluster guard, with a test asserting it) and closes a hole in a guard that already rejected-R, but it is not byte-identical GitHub behaviour, so flagging it against that intent constraint. All 20 other cases - default squash, explicit merge/rebase/--method forms, extra args, caller--sha,-d, merge failure, every other repo-override form, missing meta, malformed URLs, traversal, bad scheme, bad task id, no args - are byte-identical.bin/fm-test-run.sh tests/fm-pr-merge.test.sh- 24/24 pass, including all 13 new GitLab casesbin/fm-test-run.sh tests/fm-pr-check-security.test.sh- 33/34 pass; both GitLab cases green; 1 pre-existing unrelated failureManual end-to-end CLI transcript:bash /tmp/no-mistakes-evidence/01M0KJMZ2XCNBM4E4Q8W34X93K/gitlab-merge-demo.sh <worktree> <evidence-dir>- 11 scenarios through the realbin/fm-pr-merge.shwith a stubglabManual live-forge run (read-only, write-guarded):bash /tmp/no-mistakes-evidence/01M0KJMZ2XCNBM4E4Q8W34X93K/gitlab-live-demo.sh <worktree> <evidence-dir>against real MRs !1 and !2 ofgitlab.com/KarotKris/gitlab-merge-watch-fixtureManual GitHub parity matrix:bash /tmp/no-mistakes-evidence/01M0KJMZ2XCNBM4E4Q8W34X93K/github-parity-demo.sh <worktree> <evidence-dir> 52d20f1- 21 GitHub invocations replayed on base commit vs branch, diffed on exit status, operator output,gh-axiargv and on-disk task stateglab mr merge --help- confirmed--shahas no short alias and-Ris the only capital-R short option, so the--shaand-*R*guards have no bypassFlake isolation:test_custom_snapshot_cleanup_on_signalextracted and run 3× on agit archive 52d20f1tree and 3× on agit archive HEADtree (fails identically on both), then once with the 2s wait ceiling raised to 60s (passes)🔧 Fix: state authorised GitHub narrowing of bundled -R guard
✅ Re-checked - no issues remain.
bash tests/fm-pr-merge.test.sh- all 24 cases pass, covering the GitLab route (live-read pre-merge conditions, instance from URL, --sha binding, stale recorded head, unreadable state, glab/jq absence, head override) and the unchanged GitHub pathRan onlytest_gitlab_merge_watchandtest_gitlab_merged_poll_retiresfromtests/fm-pr-check-security.test.shin isolation (trimmed runner sourcing the worktree's real tests/lib.sh so ROOT resolves correctly) - both pass; the rest of that file was deliberately not runManual end-to-end drive of the realbin/fm-pr-merge.shthrough 18 GitLab scenarios with mockedglab/gh-axiand realjq, recording exit status, stderr, exact forge command lines, recordedpr=, and whether the merge poll was armed: happy path, self-hosted instance, each of state/detailed_merge_status/has_conflicts/blocking_discussions_resolved/pipeline-status/pipeline-sha refusing independently, no-pipeline, all-conditions-failing, stale recorded pr_head, unreadable state, glab absent, jq absent, caller--sharefused, extra flags forwarded, real merge failure propagatedGitHub parity: replayed 18 GitHub invocations (default squash,--merge,--method=merge,--rebase, extra args, caller--sha abc123,-d,--repo/--repo=/bare-Roverrides, deep org path, malformed URL, traversal, bad scheme, bad task id, missing meta, merge failure, no args) against agit archive 52d20f1export and against this branch -diff -uof the two transcripts is emptyAuthorised-narrowing evidence: same harness run on the bundled-dR wrong/repo/-yR https://other.example/g/pcluster against base and branch, showing base reachedgh-axi pr merge 6 --repo right/repo --squash -dR wrong/repoat exit 0 withpr=recorded and the poll armed, while the branch exits 1 with nothing recorded and no forge command invokedMutation/teeth check ontest_bundled_repo_override_args_refuse_before_recording: narrowed the guard from-*R*back to-Rin a scratch copy of bin/ and tests/, then ran the case with the GitHub sub-case first and again with the GitHub sub-case removed so the GitLab assertion runs first - both fail on the mutated tree (expected exit 1, got 0) and both pass on the unmutated treegit status --porcelain- worktree clean before and after; scratch repo copies, base-tree export, and sandboxes removed, evidence files retaineddocs/fm-test-portable-shards.md:23- docs/fm-test-portable-shards.md:23 and docs/fm-test-isolation-proof.json record tests/fm-pr-merge.test.sh at 6290 ms from the dated 2026-08-20 concurrency proof, and this change added ~530 lines to that proven-isolated script, so the parallel-lane balance estimates now under-count it. Left unchanged on purpose: both pages are dated empirical records of a past proof run rather than claims about the current tree, and bin/fm-test-run.sh:377 plus that doc's own contract make the durations balance hints only ("a stale hint costs a slower shard rather than lost coverage"), with the stated refresh trigger being the serial lane gaining scripts, which did not happen here. Refreshing would mean re-running the whole 24-candidate proof and rewriting timings for every candidate plus the in-code hint table, which is outside a documentation phase.✅ **Lint** - passed
✅ No issues found.
✅ **Push** - passed
✅ No issues found.