docs(verification): compaction/context regression investigation and fixes - #1
Closed
Valentino-Sole wants to merge 49 commits into
Closed
Valentino-Sole wants to merge 49 commits into
Valentino-Sole wants to merge 49 commits into
Conversation
…n unproved merge (kunchenguid#3064) * fix(pr): verify GitHub merge outcome * no-mistakes(review): Captain, fixed forge-only merge verification, queue guidance, metadata propagation * no-mistakes(document): Correct forge-specific merge documentation * no-mistakes(review): Captain: forge-only queue fix, focused tests pass * no-mistakes(review): Captain: suppress closed-state guidance and prove parent regression * no-mistakes(review): Captain: remove history proof; retain executable regressions * no-mistakes(document): Clarify GitHub recording timing in architecture docs * no-mistakes(document): Clarify outcome-aware PR merge recording documentation * no-mistakes: apply CI fixes * Revert "no-mistakes: apply CI fixes" This reverts commit c326cfa. The automatic CI repair round removed the up-front `gh` prerequisite check while keeping the `gh` dependency: `bin/fm-pr-merge.sh` still calls `gh api graphql` for the outcome read and `gh api` for the branch-rules read. That left the same hard requirement without the clear named error, and review immediately raised a new finding for exactly the failure the check prevents - `gh-axi pr merge` landing the merge while the follow-up read fails, so the PR metadata is never recorded. The check is also symmetric with the GitLab arm directly above it, which already refuses up front when `glab` or `jq` is missing, on the stated principle that a missing tool should be a named prerequisite rather than a merge that is armed and then refused for an unexplained reason. The workflows this round was chasing sit at `action_required` because this is a fork pull request; no code change can turn them green. * fix(pr): keep PR bookkeeping when a merge outcome read fails On the GitHub path a merge call that returned success was followed by `github_read_outcome || exit 1`, so a transient API failure, rate limit, or network blip during the read dropped out of the script before `record_pr_metadata` ever ran. The merge could have landed while `pr=` went unrecorded and the merge poll was never armed - bookkeeping lost on a real merge. The failure path just above already recorded metadata before exiting, so the error path was more careful than the success one. Record the PR before that refusal. Recording arms the later merge poll and is not a success claim, which is the same reasoning that keeps `record_pr_metadata` on the gh-axi failure path. The refusal itself is unchanged: exit stays non-zero and the message still names the concrete observed state. Metadata is withheld only when the read succeeds and proves the pull request neither merged nor queued. Pin it with a case that stubs `gh api graphql` into failure after a successful `gh-axi pr merge`, asserting both the non-zero exit and the recorded metadata. * no-mistakes(review): Aggregate queue rules and report conflicts explicitly * fix(pr): keep the merge abstraction reachable and its bookkeeping intact Two holes remained in the outcome-verified GitHub merge path, both on installations where gh-axi is present but gh is not. The verification preflight refused before bin/fm-pr-merge.sh ever reached the configured gh-axi merge abstraction, so an installation without gh could no longer merge at all. gh-axi now performs the merge unconditionally and the queue-aware gh read became an optional enrichment: with gh on PATH its GraphQL view still separates merged from queued, and without gh the gh-axi view still proves a landed merge while every outcome it cannot prove refuses. The PR metadata recording sat behind the outcome read, so a merge that landed before that read failed lost pr= and its merge poll. Recording now happens once, before either forge call, which arms the poll without claiming a landed outcome and leaves teardown a PR identity to verify against no matter how the read ends. Rebasing onto main also restored the durable merge-outcome reporting and the GitLab landed-state confirmation that the conflict resolution dropped. Tests pin each fix through the executable interface: the merge abstraction is reached and verified with gh absent, a failed fallback read keeps its bookkeeping, and a mock that snapshots the task meta during the forge call proves pr= is recorded before the merge can land. * no-mistakes(review): fix(pr): de-dup queue methods, fall back on failed gh read, refresh contracts * no-mistakes(review): fix(pr): quote forge output and explain armed auto-merge on refusal * no-mistakes(review): fix(pr): claim auto-merge armed only when the forge accepted it * no-mistakes(review): fix(pr): tell the operator what each GitHub refusal could not observe * no-mistakes(review): fix(pr): gate every forge-acceptance claim on a successful merge * no-mistakes(document): align merge docs with verified GitHub outcome contract
* fix(pi): stop reporting one merge to the captain twice The supervision branch's captain-outcome note told main, unconditionally, that the note "is not your own earlier output" and to relay it now. When main had already reported the same event, that assertion was false and the order turned the correct response - saying nothing new - into a mechanical re-report, so the captain saw one merge reported twice in 16 seconds. Two independent changes, both needed: - The relay instruction is now conditional. It still names itself as a supervision outcome so main cannot mistake it for its own earlier answer (the silent loss that instruction exists to prevent), and it now lets main stay quiet about an outcome it has already given the captain. - The merge case is closed at its source rather than left to that judgment. One merge reaches a home on two independent paths by design - main's own permanently main-owned merge poll, and the branch's task-local status wake - and main's captain-facing text only reaches the branch's mirror at main's turn end, so the branch can escalate before it could possibly see the captain was already told. bin/fm-pr-merge-notified.sh answers that question from bin/fm-pr-lib.sh's canonical merge-notification marker, so the answer holds regardless of mirror timing. A captain outcome naming an already-published merge is delivered as the ordinary rendered note instead of opening a follow-up turn: still appended, still visible, still recorded with the verdict the branch decided, minus the wasted turn. Any error, timeout, or unreadable state relays the outcome. A duplicate announces itself; a lost outcome does not. Regression coverage drives the real delivery path in both directions: a new outcome must still reach the captain in exactly one follow-up turn even beside an unrelated published merge, and an already-published merge must open no second turn while a different PR in the same task still does. The merge path's real producer and this new consumer are exercised end to end in tests/fm-pr-merge.test.sh. Pi-only by construction: the delivery path lives in .pi/extensions, so no other harness loads it, and the new script only reads existing markers. * no-mistakes(review): Document accepted latest-marker suppression residual * no-mistakes(review): Recheck ownership before merge outcome delivery * no-mistakes(document): Document merge-outcome suppression exception * refactor(pi): drop the source-level merge suppression, keep the envelope fix The captain reviewed this branch and judged the source-level duplicate suppression overly complicated for the problem it solved, and asked for the change to be reduced to the envelope wording alone. Remove the mergeIntoMain downgrade path, bin/fm-pr-merge-notified.sh, and every test and document that existed only for it. What remains is the conditional captain-outcome instruction: main is told to stay quiet about an outcome it has already reported and to relay anything else, which covers the duplicate without a second mechanism. The silent-loss protection is untouched - the note is still typed, self-describing, and delivered as one invisible follow-up turn - and the behavioral tests still assert that, now requiring both halves of the conditional instruction. * no-mistakes(ci): Clarified in code comments and owned documentation that this is intentionally an M1-only, model-facing conditional relay fix—not source-level suppression—addressing Greptile’s mistaken scope expectation without changing runtime behavior. Net diff remains 3 files and 27 insertions. Verified with fm-pi-branch-extension tests, fm-lint, doc audience check, and git diff --check; all passed * no-mistakes(ci): Strengthened the runtime delivery test to verify the captain outcome retains its required self-description and outcome text. Verified with `bash tests/fm-pi-branch-extension.test.sh`, `bin/fm-lint.sh`, `bin/fm-doc-audience-check.sh`, and `git diff --check`; all passed. The outer pipeline can now commit and attest the new head
* fix(bin): bind the live pipeline-owned run instead of a superseded failed row fm-crew-state.sh bound a superseded FAILED no-mistakes run to a task instead of the LIVE replacement run: the live run's pipeline-owned lane head is not a git object in the task worktree, so head-equality attribution rejected it and the coarse runs-list fallback silently continued past the RUNNING row onto an older failed row whose head equalled the stale worktree HEAD. The home summary then flipped invalid and Bearings hid the home's live work (F10). Attribution precedence now follows the daemon's own identity: - An ACTIVE run for the task's branch binds without head equality while branch_sync.state is pipeline_owned (fm_nm_run_is_pipeline_owned_active); the pipeline owning the branch is itself the attribution. - A genuinely failed run with no later run on the branch still reports failed through the unchanged head-equality path - real failures are not hidden. - In the coarse runs scan, an unresolvable head is unknown attribution and stops the scan (fm_nm_head_resolvable) instead of falling through to an older row; a resolvable-but-mismatched head keeps the historical reused-branch skip. The exemption never applies to a terminal run and requires pipeline_owned specifically, both pinned by negative-control tests. Fixture shape verified against the live incident run's real axi status output. * no-mistakes(document): Updated run-attribution documentation ownership
…unchenguid#3211) * fix(pi): surface requested supervision outcomes * no-mistakes(review): Mirror in-flight captain requests before branch dispatch * no-mistakes(review): Exercise real branch ownership and main outcome access * no-mistakes(review): Preserve request tails and align verdict guidance * no-mistakes(review): Preserve complete current captain requests * no-mistakes(review): Require visible requested outcomes and realistic classification * no-mistakes(document): Align supervision outcome documentation * no-mistakes(ci): Fixed Greptile’s runtime-ordering finding. The extension now stages Pi’s authoritative `before_agent_start` prompt before SessionManager persistence and suppresses the later duplicate entry. Updated docs and behavioral regression to reproduce real Pi ordering and verify each prompt is mirrored exactly once. Passed branch-extension tests, supervision tests, strict Pi typecheck, full lint, and diff checks * no-mistakes(review): Use canonical operational input classification * no-mistakes(review): Filter legacy operational inputs canonically * no-mistakes(document): Clarify captain request mirroring boundary * no-mistakes(ci): Fixed the CI time-boundary failure in tests/fm-public-followup.test.sh by pinning its clock, including context-registry setup. This prevents follow-up fixtures from expiring based on wall time. Verified the full regression suite passes, project-owned lint passes, and git diff checks are clean * no-mistakes(document): Clarify captain-visible supervision outcome documentation
…#3210) * feat(bin): per-home remote transport lanes with cancellation, bounded send, and closed stdin All remote commands for every home on one host used to serialize through one single-job-at-a-time worker on one shared queue: a timed-out caller abandoned a staged job that kept running, retries convoyed behind it, fm-send's remote leg had no time bound, and staging captured the caller's stdin to EOF so any fm-on.sh caller with an open stdin wedged staging indefinitely. - The worker now serves one lane per staged home: same-home jobs run strictly FIFO in a new staging-sequence order while different homes run concurrently, each lane as its own top-level worker process (a backgrounded subshell does not reliably reap dead children, so a zombie group leader kept a finished command's process group signalable). Long-poll preemption is lane-scoped. - A caller that disconnects or times out cancels its job: the entrypoint marks the record on any post-staging exit and probes its parent so a dead ssh channel cancels without a signal; the worker skips cancelled queued jobs, terminates a running cancelled job's process group, and reaps the record. - fm-send's remote leg is bounded by FM_SEND_REMOTE_BUDGET (default 30s) and a bound hit exits through the existing unconfirmed-delivery contract, which stays idempotent because the remote enqueue deduplicates. - fm-on.sh defaults the remote command's stdin to /dev/null; the three payload callers pass the new --stdin flag. Abandoned .stage.* litter is age-reaped. - The job execution deadline no longer loses up to a second to clock truncation. * no-mistakes(review): Protect live stages and validate send budgets early * no-mistakes(review): Preserve sequence lock ownership during stale recovery * no-mistakes(review): Allocate job sequences at publication boundary * no-mistakes(review): Bound remote keys and extend stale lock recovery * no-mistakes(document): Document bounded remote transport behavior * no-mistakes(lint): Suppress intentional deferred-expansion lint warning * no-mistakes(ci): Fixed stale sequence-lock recovery by reconciling the counter against published job records before allocating the next sequence, preventing duplicate sequences and same-home FIFO violations. Added a behavioral regression test reproducing displacement after publication and verifying execution order. Passed fm-remote-transport-lanes.test.sh, fm-remote-job.test.sh, fm-lint.sh, and git diff --check * no-mistakes(review): Use atomic sequence claims and lossless lane keys * no-mistakes(review): Recover regressed sequence hints and rate-limit claim reaping * no-mistakes(review): Restrict worker heartbeats to serving loop * no-mistakes(review): Verify supervisor identity before lane recovery signals * no-mistakes(review): Verify tracked lane and claim owner identities * no-mistakes(document): Clarify remote lane and transport contracts * no-mistakes(ci): Fixed the CI time-boundary failure by pinning fm-public-followup tests to a deterministic clock, including context-registry setup. Verified tests/fm-public-followup.test.sh, tests/fm-remote-transport-lanes.test.sh, shellcheck, and git diff --check * no-mistakes(review): Preserve assigned lane ownership of queued jobs * no-mistakes(review): Reserve homes owned by foreign queued lanes * no-mistakes(review): Preserve completed results during crash recovery * no-mistakes(review): Harden claim cleanup, expiry, and cancellation races * no-mistakes(review): Verify process groups and reap abandoned results * no-mistakes(review): Stop leaderless groups and reap cancelled publications * no-mistakes(document): Correct remote transport lifecycle documentation * no-mistakes(lint): Quote done state comparisons for ShellCheck
* fix(tests): make the changed-file map select per script and stabilize a budget flake
The changed-file map's bin/ fallback resolved a direct test reference to that
test's whole FAMILY. bin/fm-push-transition-lib.sh is named by exactly one
real-Herdr E2E, so a one-line change to it selected all 12 real-herdr-gated
scripts, including a 341s presentation E2E with no dependency on it.
Resolve direct test references per script, and keep resolving consumer bin/
scripts through the curated map so recorded family-level coupling survives.
Also fix a load-sensitive flake: the tool-update budget deadline is whole-second
granular, so a test budget of 1 left headroom anywhere in (0, 1] seconds and the
first budget check could already read as exhausted.
* feat(bin): make suite wall clock a result and let a family's concurrency be proven
--max-wall-ms fails a run whose wall clock exceeds the caller's budget, after
reporting the per-script results. A suite that stays green while outgrowing its
caller's invocation budget is the regression that got an agent killed mid-run
and retried invisibly, so duration has to be a result rather than a log note.
--pool on the isolation-proof harness runs the same concurrent proof over a
whole family, so 'is this family safe to parallelize?' is answered by a command
instead of a guess. Measured watcher-wake-lock and refused it: 3 of 18 scripts
fail under concurrency on wall-clock assertions about reaching the next poll.
* perf(bin): schedule the changed suite concurrently, longest first
The watcher-wake-lock family is proven concurrent-safe (two clean runs, 18
candidates, 0 failures at 4 workers; docs/fm-test-isolation-proof.md), so
--changed now schedules its proven-concurrent scripts with bounded parallelism
and runs any unproven remainder serially afterwards, never beside them.
Concurrent runs are ordered longest-hint-first. Workers are handed scripts in
order, so alphabetical order started the 193s fm-watch-triage last and stranded
it running alone: 395s wall against a 205s balanced four-worker sum.
An explicit --jobs keeps its strict refusal, so every CI lane is unchanged.
* fix(bin): bound a hung test instead of letting it hang the suite
tests/fm-calm-pi-extension.test.sh was observed running 17+ minutes against a
464ms recorded hint, and the suite had no per-script bound to stop it. An
unbounded suite is precisely what silently outruns a caller's invocation budget,
and --max-wall-ms is evaluated after the run so it cannot end one that never
finishes.
--per-script-timeout-secs terminates a script that outruns it and records exit
124, so the run still completes, accounts for the script, and fails. The
auto-concurrent --changed path applies 900s, far above the slowest real script
(the 341s Herdr presentation E2E), so it only ever converts a hang.
* no-mistakes(review): Enforce safe concurrency and descendant timeouts
* no-mistakes(review): Validate empty runs and isolation proof pools
* no-mistakes(review): Measure selection time in wall budget
* no-mistakes(review): Reap interrupted workers and bound finalization
* no-mistakes(review): Contain shutdown descendants and watchdog finalization
* no-mistakes(review): Honor remaining budget and close launch races
* no-mistakes(review): Restore timeout helper and simplify runner cleanup
* no-mistakes(review): Record isolation pool admission metadata
* no-mistakes(review): Bound Chrome reap and scope proof admission
* no-mistakes(review): Align proof scheduling and preserve budget summaries
* no-mistakes(review): Remove unreliable finalization watchdog
* no-mistakes(review): Freeze budget duration and enforce admission caps
* no-mistakes(document): Refresh test runner concurrency documentation
* no-mistakes(lint): Fix ShellCheck findings in test runner scripts
* no-mistakes(ci): Fixed Greptile’s concurrency-consent finding. `--changed` now remains serial by default; `--changed --jobs auto` explicitly opts into bounded concurrency and the automatic hang timeout. Updated documentation and added behavioral coverage proving serial default behavior, explicit concurrent scheduling, and refusal of `--jobs auto` outside `--changed`. Verified with `bash tests/fm-test-run.test.sh`, `bin/fm-lint.sh`, and `git diff --check`
* no-mistakes(review): Restore automatic changed-suite concurrency and timeout
* no-mistakes(review): Correct changed-suite contributor guidance
* no-mistakes(review): Reject gate-skipped isolation proofs
* no-mistakes(review): Correct automatic concurrency evidence
* no-mistakes(review): Isolate nested runner process groups
* no-mistakes(review): Remove unreliable signal cleanup machinery
* no-mistakes(test): Narrow changed-suite selection to executable contract owners
* no-mistakes(document): Document isolation proof skip and artifact semantics
* no-mistakes(ci): Fixed Greptile’s concurrency-consent finding. `--changed` now remains serial by default; bounded concurrency requires explicit `--jobs auto`. Updated behavioral coverage, contributor guidance, and isolation-proof commands accordingly. Verified with `tests/fm-test-run.test.sh`, `bin/fm-doc-audience-check.sh`, `bin/fm-lint.sh`, Bash syntax checks, and `git diff --check`; all passed
* no-mistakes(review): Restore plain changed-suite automatic concurrency
* no-mistakes(review): Record resolved changed-suite worker count
* fix(bin): keep a runner change selecting its whole curated family
A pipeline fix round narrowed the curated changed-file map so bin/fm-test-run.sh
and bin/fm-test-isolation-proof.sh selected only their own two contract tests,
and the documentation surfaces only the audience test. That cut this branch's
own changed selection from 33 scripts to 5.
The runner executes every pure-contract-unit script, so its contract test
passing proves its logic is right, not that the suite it drives still runs.
Narrowing it also makes any wall-clock claim about the changed suite trivially
true by not running the work.
Only the unmapped bin/* grep fallback resolves per script; curated mappings keep
their recorded family coupling.
* perf(bin): admit the pure-contract-unit family to bounded concurrency
A runner-file change selects pure-contract-unit, so that family decides the
changed suite's wall clock. With only watcher-wake-lock admitted, 14 of its 33
selected scripts fell to the serial tail and the selection measured 327.3s
against a 300s budget: the concurrent group was 19 scripts totalling 273.4s
while the tail alone was 215.7s.
bin/fm-test-isolation-proof.sh --pool pure-contract-unit --jobs 4 passes twice,
32 candidates, 0 failures, so the family is admitted on recorded evidence.
Full 33-script plain --changed: 327.3s -> 181.8s / 178.5s / 172.7s, 0 failures,
inside a 300000ms budget. Also states the per-script guard's derivation.
* no-mistakes(review): Align contract-unit concurrency cap with recorded proof
* no-mistakes(document): Record final changed-suite performance evidence
* fix(bin): keep an empty changed selection clean on stock macOS Bash
Under set -u, bash 3.2 treats "${arr[@]}" on an EMPTY array as an
unbound-variable error, while bash 4.4+ makes it a harmless no-op. The
concurrency work removed the early exit for an empty selection, so execution
fell through to the unguarded existence loop: on stock /bin/bash 3.2.57 a
contributor who changes only documentation and runs --changed got
bin/fm-test-run.sh: line 1713: SCRIPTS[@]: unbound variable
with exit 1 and no summary, instead of a clean total=0 pass.
Restore the early exit, and guard every remaining array expansion reachable
with an empty selection. The reported duration is real elapsed invocation
time rather than a hardcoded zero, so a selection phase that outran
--max-wall-ms still fails.
Verified on this host with /bin/bash 3.2.57: exit 1 with the unbound-variable
error before, exit 0 with FM_TEST_SUMMARY total=0 after.
* no-mistakes(document): Document shell-bound changed-suite performance
---------
Co-authored-by: Kun Chen <kun-1@kunchenguid.com>
* feat(bin): publish per-home summary ledger * no-mistakes(review): Bound and schedule home summary publication * no-mistakes(review): Prove recurring watcher summary refresh cadence * no-mistakes(review): Bound refresh workers and publish durable spawns * no-mistakes(review): Fix atomic kill process-group coverage * no-mistakes(review): Bound state initialization within refresh timeout * no-mistakes(document): Document recurring bounded home-summary publication * no-mistakes(review): Bound and log all best-effort refresh failures * no-mistakes(review): Harden cadence and timeout regression coverage * no-mistakes(document): Document home-summary runtime tuning * no-mistakes(lint): Fix direct exit-code check in refresh test * no-mistakes(ci): Fixed remote secondmate retirement recreating the deleted home: teardown now skips side-band summary refresh when its overridden state directory was removed. Verified with remote lifecycle E2E, teardown tests, home-summary tests, ShellCheck, and git diff checks * no-mistakes(document): Clarify atomic home-summary publication guarantee
* fix(pi): gate first call on startup context * no-mistakes(document): Correct Pi startup prerequisite verification date * no-mistakes(review): Captain, fix startup process-group retirement after leader exit * no-mistakes(review): Captain, release reload exit listeners on shutdown * no-mistakes(review): Captain, complete startup exit lifecycle ownership * no-mistakes(review): Captain, release empty startup process-group ownership promptly * no-mistakes(review): Captain, supervise startup ownership and restore failure fallback * no-mistakes(review): Captain, restore live Pi supervisor execution * no-mistakes(document): docs: clarify Pi startup prerequisite delivery
* fix(pi): restore 0.84.4 adapter compatibility * no-mistakes(review): Restore Pi collapsed and expanded outcome parity * no-mistakes(review): Preserve Pi stock previews through capability probing * no-mistakes(document): Document Pi 0.84.4 renderer compatibility
…nchenguid#3273) * fix(bin): keep home-summary publication bounded and off the watcher beat A home whose tasks had accumulated ordinary status history could not publish state/home-summary.json at all, and every attempt starved the watcher's liveness beacon while it failed silently. The producer's per-task open-decision fold spent tens of milliseconds per status line on a bash 3.2 global bracket-class substitution used only as a blank-line guard. On a real home that made the whole ledger producer take minutes, so publication burned its full FM_HOME_SUMMARY_TIMEOUT on every attempt and never completed. Replace that guard with an equivalent case glob in the one fold owner, which both the whole-file and cursor-backed folds use. Bound each per-task current-state read in the snapshot with FM_SNAPSHOT_CREW_STATE_TIMEOUT. For a remote secondmate that read crosses ssh, whose dead-peer detection deliberately never kills a slow-but-alive remote command, so nothing else bounded it. Detach the watcher's two publication triggers from the poll loop. The loop owns the beacon that fm-guard.sh reads as proof supervision is alive, and an inline publication put up to a full publication deadline between two beacon touches. A single in-flight publication is tracked so a slow one cannot accumulate clones. Report a repeatedly failing publication at session start. Publication stays deliberately non-fatal to its caller, so the existing bounded home-local failure record is now surfaced as a HOME_SUMMARY bootstrap line once the ledger is absent or stale and failures have been recorded since. * no-mistakes(review): Preserve home-summary failure attempt ordering * no-mistakes(review): Enforce durable home-summary single-flight and ordering * no-mistakes(review): Derive failure ordering from publication boundaries * no-mistakes(review): Restore best-effort failure logging and publication scoping * no-mistakes(review): Make ordering regression sensitive to one failure * no-mistakes(document): Correct HOME_SUMMARY diagnostic guidance
…henguid#3268) * fix(supervision): classify the appended status span, not the last line An actionable project update could be classified as routine and absorbed, so a worker that raised a decision, hit a blocker, failed, or finished stalled silently with the captain never told. Trigger, mask, symptom. A worker appends a captain-relevant event (`needs-decision`, `blocked`, `failed`, `done`). Any later routine append - a `working:` progress note - lands before the supervisor classifies the batch; the watcher's 30s signal-grace linger exists precisely to coalesce a status write with the same turn's turn-end, so this window is ordinary rather than rare. Both supervisors then asked "is the LAST line captain-relevant?", read the routine line, and absorbed the wake. The `.seen-*` suppressor advanced either way, so nothing ever re-read the event. When the crew was also provably working, the no-verb fallback absorbed it too, which is why the event disappeared completely instead of surfacing late. Reproduced end to end against a real watcher before any change: with the trailing `working:` append the watcher never exits and the wake queue stays empty; with that one line removed - the smallest counterfactual - the same `needs-decision` surfaces and queues. The away-mode daemon's `classify_signal` returns `self|routine signal` for a `blocked:` event under the same mask, which is the worse case because no captain is present to notice. The proven path was already in the tree: `status_open_decisions` fixed this exact masking for the durable decision fold, and its header states the rule - reading an append-only event log last-event-wins cannot represent an earlier event that a later unrelated line moved past. The classification path was never migrated to that read model. That is the earliest divergence, and the fix is to migrate it rather than to special-case the symptom. `status_span_first_actionable` in bin/fm-classify-lib.sh is the new single owner: it reads the bytes at or after a caller-supplied position and returns the first still-live captain-relevant event. Each supervisor supplies its own position, because the always-on watcher and the away-mode daemon classify the same stream independently and must not share one cursor: the watcher reads the size already recorded in its `.seen-*` signature (no new state) and its `.hb-surfaced-<task>` backstop marker, and the daemon its `.subsuper-seen-status-<task>` marker. Those two markers held the escalated line and now hold the escalated-through byte offset, which also removes a second defect in the same code - content dedup silently swallowed a genuinely new event whose text repeated an older one. An absent, malformed, or past-the-end position reads the whole log, so uncertainty surfaces events rather than losing them, and a marker an older build wrote as a status line reads that way too. Status logs are only ever appended to, including across a reused task id, so a recorded position keeps its meaning. A `needs-decision`/`blocked` event in the span is retired only when the whole-file fold proves its key closed; `status_open_decisions` stays the sole owner of that rule, so same-key reopening and reserved-key namespaces need no second implementation here. Every other captain-relevant event is terminal and always actionable. Both backstops now walk every status log instead of only those whose last line looks captain-relevant, because the event a backstop most needs to catch is exactly one a later append has moved past. That leaves `scan_captain_relevant_statuses` with no callers, and it is removed rather than left as a working copy of the defective read model. Regression coverage exercises the classifier and both supervisors through their own interfaces: the masked decision, the captain-reported release/install completion followed by cleanup chatter, and the away-mode blocker all surface; a routine append after an already-classified event stays absorbed, so the fix does not convert ordinary progress into wakes; and the heartbeat backstop catches a masked event the per-wake path missed. The end-to-end watcher tests drive a real fm-watch.sh with the crew reported as provably working, which is the configuration that made the original stall silent. Two further claims in the supplied RCA are deliberately not patched here. "Repeated operational recoveries produced all-clear replies despite known actions" is downstream of this same cause, not an independent contributor: an all-clear reply is the documented response when the specific event needs no action, so a classification that wrongly reported "no action" produces it, and correcting the classification removes it. "The project was subjected to validation requirements outside its accepted path" is delivery-mode selection, which AGENTS.md section 7 owns; no code changed here touches it, so it is out of scope. Harness and backend axes were inspected rather than assumed: nothing in this path reads a vendor-emitted signal. The status log's format and append protocol are Firstmate's own and identical for every harness, and no runtime backend reads or writes `.status` files (`bin/backends/*` contain no reference to them). The surrounding triage's only backend touchpoints - pane capture and the authoritative crew-state read - are unchanged. No live-harness guard applies and no per-harness verification record changes. Verified with `bin/fm-lint.sh`, `bin/fm-doc-audience-check.sh`, and `bin/fm-test-run.sh --changed --base origin/main`. * no-mistakes(review): Prevent status races and surface classification failures * no-mistakes(review): Surface unreadable signals and preserve AFK endpoints * no-mistakes(review): Route stale wakes through captured span verdicts * no-mistakes(review): Retire supervision offsets with reused task state * no-mistakes(review): Bind status offsets and preserve live decision origins * no-mistakes(review): Strengthen status identity with verified birth time * no-mistakes(review): Skip turn-end markers during status classification * no-mistakes(review): Preserve status presentation with platform-strength identities * no-mistakes(review): Retain failed wakes and advance routine checkpoints * no-mistakes(review): Surface all events and retain unreadable wakes * no-mistakes(review): Treat absent status logs as successful empty spans * no-mistakes(review): Bound repeated classification failures with durable receipts * revert(supervision): drop the failure-receipt and durable-retry machinery Captain-authorized revert to the minimal fix. Review rounds added a durable failure-receipt store and wake-retention-on-failure to bound repeated classification failures. That machinery grew larger than the fix it protected and kept producing its own defects: an unreadable log still looped forever because the always-on watcher never consulted the receipt, and the receipt was persisted before its diagnostic was durably queued, so a crash in between swallowed the alarm outright. Those two defects go away with the code that contained them rather than being repaired. Removed: the failure-receipt path, fingerprint, record and clear helpers and their retirement bookkeeping; the retention of a durable wake when classification fails; and the error-propagation plumbing in both supervisors that existed only to drive them. Kept, because it is the accepted fix rather than the declined machinery: span classification of the events appended since a supervisor last looked, in both supervisors and both backstops; reporting every actionable event in a span and committing a position only through what was reported; naming the live opening of a reopened decision; treating an absent log as ordinary and an unreadable one as worth reporting; the non-.status filter; and the platform-strength identity that guards a position commit without failing a read. Replacement behavior for a log that cannot be classified: report it once, do NOT advance the classification position so the content is classified from where it stopped once readable, and DO advance the wake signature so the report is bounded to one per distinct file state. Reporting and reading are different acts: telling the captain about a log is not the same as having read it, and only the latter may move a classification position. The residual risk is explicit and accepted: there is no guaranteed automatic retry inside a crash-mid-read window, and the locked session-start replay of the durable queue covers it. That rationale is recorded at mark_escalated_seen so a future reader does not reintroduce the retry as a "missing" guarantee. Also fixes lint failures that arrived with the review-fix commits and were never caught because the run never reached its lint step: an unfollowable conditional source directive, a second unquoted-expansion site left after a call was split across lines, cleanup of the file being read inside its own read loop (restructured to one post-loop teardown rather than three in-loop copies), stub functions in tests that are invoked indirectly, and a test local left unused when its assignment was replaced by a helper. bin/fm-lint.sh passes on the default branch, so these were introduced here. Verified with `bin/fm-lint.sh`, the end-to-end masked-decision and away-mode reproductions, and `bin/fm-test-run.sh` over the supervision, wake-queue, wake-drain, watch-arm and inactive-reconcile suites (6 scripts, 0 failures). * no-mistakes(review): Correct classification failure contract documentation * no-mistakes(review): Bound unreadable status reports without skipping classification * no-mistakes(review): Preserve escalation markers when buffering fails * no-mistakes(review): Detect permission recovery without advancing classification * no-mistakes(document): Document status span classification contract * no-mistakes(ci): Fixed CI failures by lazily loading classification helpers in fm-wake-lib, preserving minimal recovery/remote fixtures; added a public current-status marker helper and updated behavioral fixtures to use the v2 marker contract; resolved ShellCheck variable collisions in fm-control and fm-public-followup-lib. Verified fm-lint, bash syntax, fm-control, public-followup, wake-queue, send-resolve-key, captain-hold, pending-reply, remote-reply, remote-backlog-handoff, turnend-guard, and Claude autoarm tests. The Pi branch suite reached a separate local stock-render mismatch under Node 24; its CI-reported missing-classifier failure path is fixed * no-mistakes(review): Escalate blockers while preserving declared-wait cadence * no-mistakes(review): Clarify actionable events override wait self-handling * no-mistakes(review): Surface rejected decisions and dangling status links * no-mistakes(document): Document reserved-key reconciliation classification * no-mistakes(ci): Fixed the flaky portable serial CI test by modeling the retained staging directory as genuinely owned by a live process and aging both fixtures deterministically. This removes scheduler-timing dependence while verifying the worker reaps abandoned staging and preserves live staging. Verified with fm-remote-transport-lanes.test.sh, bin/fm-lint.sh, bash syntax, and git diff --check * no-mistakes(document): Correct away-mode classification documentation
…#3289) * docs: split harness adapter operations reference * no-mistakes(review): Fix harness adapter routing and ownership contracts * no-mistakes(review): Prune duplicate harness adapter ownership prose * no-mistakes(review): Fix default effort routing and Grok max semantics * no-mistakes(review): Remove source-only routing test and duplicate semantics * no-mistakes(review): Add local harness adapter instruction evaluation * no-mistakes(review): Fix harness evaluation gating and change mapping * no-mistakes(test): Captain, require explicit harness instruction evaluator model * no-mistakes(document): Fix harness adapter documentation references
* test(fixtures): share fake-toolchain and spawn-world builders Future tests can start from tests/fixtures.sh instead of copying stubs, and a no-mistakes version-floor bump is one constant rather than a multi-file edit. Migrated this round: fm-busy-adapter-wiring, fm-spawn-pool-base-freshen, fm-grok-harness, fm-tangle-guard, fm-gate-refuse, fm-spawn-dispatch-profile. Left for opportunistic migration: remaining make_spawn_fakebin copies (trace-context, kimi, muse, backend), the make_stubs send cluster, and the fake no-mistakes version banners in bootstrap/session-start/secondmate suites. Did not touch tests/fm-pr-check-security.test.sh. * no-mistakes(review): Prevent fake SSH test from blocking on stdin * no-mistakes(document): Clarify shared fixture documentation * no-mistakes(ci): Fixed the flaky watcher triage test by extending its startup-sensitive timer-repair wait from 3s to 10s, matching existing loaded-runner budgets. Verified with the full tests/fm-watch-triage.test.sh suite, bash syntax validation, and git diff checks * no-mistakes(ci): Fixed portable serial shard 4 by updating the inactive-reconcile fixture to prime status through the public fm_wake_status_mark_current API, ensuring classifier helpers load correctly and preventing the idle watcher from exiting. Verified the test three consecutive times, ran fm-test-fixtures, ShellCheck, bash syntax checks, and git diff checks. The outer no-mistakes executor can now bind a fresh attestation to the new head * no-mistakes(ci): Added behavioral coverage proving the shared spawn tmux fixture defaults an unset FM_FAKE_PANE_PATH to empty. Verified the fixture suite, ShellCheck, syntax/diff checks, and all six migrated test suites; all passed. The outer executor can now bind a fresh no-mistakes attestation to the updated head
* feat(bin): retire completed PR-check migration machinery Every registered home already carried both completion markers, and no installer still creates pre-migration checks. Remove the one-time migrate script, its bootstrap/watch/teardown/docs surface, and migration-path tests without weakening live check-trust or PR-poll authentication. * no-mistakes(review): Restore live PR-check security coverage * no-mistakes(document): Refresh retired PR-check documentation * no-mistakes(ci): Fixed both failing CI checks. Updated inactive-reconcile setup to use the public status-marking interface, preventing false watcher exits. Made remote-job shutdown deterministic by stopping the complete worker tree before tampering. Verified both affected test suites, repeated inactive reconciliation, shell syntax, and git diff checks
…3247) * feat(extensions): bind trusted external process-event adapters * no-mistakes(review): Enforce owner and remote-home conformance * no-mistakes(review): Enforce serialized remote extension package lifecycle * no-mistakes(review): Enforce identity-conditional extension retirement * no-mistakes(review): Serialize extension retirement and recover crash cuts * no-mistakes(review): Unify retirement worker and lifecycle lock ownership * no-mistakes(review): Harden extension lifecycle retirement serialization * no-mistakes(review): Unify extension registration and overridden-state lifecycle boundaries * no-mistakes(document): Clarify built-in-only captain answer routing * no-mistakes(lint): Captain: fix extension binding ShellCheck findings * 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 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Use isolated UID mapping for owner conformance * no-mistakes(review): Captain: remove forbidden CI ownership wrapper * no-mistakes(review): Serialize extension binding publication * no-mistakes(review): Document ordinary CI owner-fixture exclusion * no-mistakes(review): Quarantine orphaned handshake descendants * no-mistakes(test): Fix orphan attribution * no-mistakes(test): Harden process tracker baseline * no-mistakes(test): Harden detached descendant attribution * no-mistakes(test): Use exact invocation-group cleanup * no-mistakes(test): Bound remote conformance transport crossings * no-mistakes(test): Parallelize isolated extension conformance tests * no-mistakes(test): Lifecycle suite still exceeds deadline * feat(extensions): bind trusted external process-event adapters * no-mistakes(review): Enforce owner and remote-home conformance * no-mistakes(review): Enforce serialized remote extension package lifecycle * no-mistakes(review): Enforce identity-conditional extension retirement * no-mistakes(review): Serialize extension retirement and recover crash cuts * no-mistakes(review): Unify retirement worker and lifecycle lock ownership * no-mistakes(review): Harden extension lifecycle retirement serialization * no-mistakes(review): Unify extension registration and overridden-state lifecycle boundaries * no-mistakes(document): Clarify built-in-only captain answer routing * no-mistakes(lint): Captain: fix extension binding ShellCheck findings * 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 * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Use isolated UID mapping for owner conformance * no-mistakes(review): Captain: remove forbidden CI ownership wrapper * no-mistakes(review): Serialize extension binding publication * no-mistakes(review): Document ordinary CI owner-fixture exclusion * no-mistakes(review): Quarantine orphaned handshake descendants * no-mistakes(test): Fix orphan attribution * no-mistakes(test): Harden process tracker baseline * no-mistakes(test): Harden detached descendant attribution * no-mistakes(test): Use exact invocation-group cleanup * no-mistakes(test): Bound remote conformance transport crossings * no-mistakes(test): Parallelize isolated extension conformance tests * no-mistakes(test): Lifecycle suite still exceeds deadline * no-mistakes(review): Split extension conformance and forward remote transfer input * no-mistakes(review): Forward malformed remote payloads through fm-on * no-mistakes(review): Bound extension coordinator failure cleanup * no-mistakes(test): Skip repeated orphan sweep in coordinator children * no-mistakes(test): Queue isolated extension sections through bounded workers * no-mistakes(test): Bound extension coordinator lane cleanup * no-mistakes(test): Split remote lifecycle coordinator sections * no-mistakes(test): Coordinator probes pass; aggregate deadline remains * no-mistakes(test): Launch extension sections concurrently * no-mistakes(test): Fix coordinator marker publication * no-mistakes(test): Stabilize extension binding coordinator timing * no-mistakes(lint): Fix extension binding ShellCheck warnings * fix(extensions): prove invocation cleanup before retirement * no-mistakes(review): Harden process-event inbox confinement * no-mistakes(review): Preserve legacy capture parity * no-mistakes(review): Protect external registry staging * no-mistakes(test): Stabilize bounded extension conformance aggregate * no-mistakes(document): Document external evidence confinement * no-mistakes(ci): CI phase fixed. The failure was a flaky fixture in `tests/fm-remote-transport-lanes.test.sh`: its “fresh/in-use” staging directory had no live owner identity, so the real worker correctly reaped it once the 1-second age boundary elapsed on slower CI. The fixture now records the active test shell’s exact PID/start identity and cleans those records before removal. Verified: `bash tests/fm-remote-transport-lanes.test.sh` exits 0 with all checks passing; `git diff --check` passes. Provider check retrieval was also retried successfully, resolving the selected manual CI finding. Changed file: `tests/fm-remote-transport-lanes.test.sh` * no-mistakes(review): Harden extension staging and lifecycle reservation * no-mistakes(review): Harden external staging and lifecycle reservations * no-mistakes(review): Wire capture helper into remote conformance * no-mistakes(review): Pin external capture handoff and signal failures * no-mistakes(review): Bind pinned capture authority to inherited descriptor * no-mistakes(review): Harden descriptor-bound capture authority * no-mistakes(review): Harden core capture reservation authority * no-mistakes(review): Harden capture reservation boundaries * no-mistakes(review): Harden capture reservations and cleanup * no-mistakes(review): Harden capture handoff and reservation cleanup * no-mistakes(review): Bind capture handoff to claim descriptors * no-mistakes(review): Release lifecycle locks after host crashes * no-mistakes(review): Pin reservation recovery to recorded state roots * no-mistakes(review): Reject control bytes in claim state roots * no-mistakes(test): Stabilize extension capture descriptor handoff * no-mistakes(document): Document extension capture authority boundary * no-mistakes(lint): Fix ShellCheck extension binding warnings * no-mistakes(ci): CI phase result: fixed `bin/fm-procevent.sh` by initializing the shared `capture_state` sentinel for built-in adapters under `set -u`. This prevents normal built-in captures from aborting before publication. Verified: `bash -n bin/fm-procevent.sh` and `git diff --check` pass. The focused process-event suite was run locally but stopped earlier at a local detached-runner claim failure (`reconcile never claimed the registered source`), before the CI-reported post-capture path; CI evidence confirms the fixed unset-variable failure affected the failing remote, board, watcher, and process-event checks * no-mistakes(document): Correct extension namespace creation timing * no-mistakes(lint): Initialize capture locals for ShellCheck
* fix(bin): deliver the real definition of done to a promoted scout, and ban --yes A promoted scout used to receive a free-form placeholder instead of the mode-specific Definition of done a briefed ship worker gets, so it never saw the ask-user escalation rule or the --yes prohibition. That gap is the concrete reason one incident's worker drove validation with --yes and answered its own ask-user findings. - Add bin/fm-dod-lib.sh as the single owner of a ship task's mode-specific Definition of done, rendered by both bin/fm-brief.sh and bin/fm-promote.sh so the two contracts cannot drift. - bin/fm-promote.sh now writes data/<id>/ship-instructions.md carrying the scratch inventory, clean base, ship branch, and that Definition of done, and prints the fm-send.sh command that delivers it. - State the --yes ban as a prohibition rather than a preference, without claiming an enforcement the tool does not provide. - Cover both through the real promotion and brief paths in tests/fm-task-delivery.test.sh and tests/fm-brief.test.sh. * no-mistakes(review): Publish promotion instructions before committing task state * no-mistakes(review): Supersede conflicting scout delivery rules after promotion * no-mistakes(review): Reject invalid promotion instruction destinations * no-mistakes(document): Align documentation with promotion delivery contracts * no-mistakes(ci): Fixed both CI findings. Promoted workers now receive an explicit worktree-isolation check before branch creation, with instructions to stop and escalate if they are in the primary checkout. Updated behavioral coverage to verify the delivered promotion payload, and aligned the ask-user authority test with the new fleet-wide --yes prohibition. Verified with bin/fm-lint.sh, tests/fm-brief.test.sh, tests/fm-ask-user-authority.test.sh, tests/fm-task-delivery.test.sh, and git diff --check * no-mistakes(ci): Made tests/fm-ask-user-authority.test.sh executable so the modified colocated behavioral test runs directly like the surrounding test suite. Verified bin/fm-lint.sh, fm-brief, ask-user-authority, and task-delivery tests; all pass. git diff --check is clean * no-mistakes(ci): Strengthened tests/fm-task-delivery.test.sh to behaviorally verify that real promotion and brief generation deliver byte-identical Definition-of-done blocks for all three modes. Verified tests/fm-task-delivery.test.sh, tests/fm-brief.test.sh, bin/fm-lint.sh, and git diff --check. The outer pipeline can now commit and attest the updated head * no-mistakes(ci): Fixed promotion isolation instructions so any checkout other than the launched disposable worktree requires escalation, including another non-primary worktree. Updated behavioral coverage against the delivered promotion payload. Verified fm-task-delivery, fm-brief, fm-ask-user-authority, full fm-lint/ShellCheck, workflow lint, and git diff checks
) * fix(bin): present complete Lavish board feedback as structured output Give the Lavish adapter a read-only presentation so a handler sees every annotation and the session-ending tag=message as its own field, instead of grepping a truncated raw capture. * no-mistakes(review): Preserve unquoted messages and prioritize captain prose * no-mistakes(document): Document structured Lavish result reads * no-mistakes(ci): Fixed Lavish `read` completeness: rows missing declared fields are excluded from presented items, counted as malformed, and force `complete: no`. Added behavioral regression coverage through the adapter interface. `bin/fm-lint.sh`, syntax checks, and focused valid/malformed read checks passed. The portable-serial failure was an unrelated secondmate cooldown timing flake
* fix(records): pair backlog transitions with the record that moves Dispatch and completion each moved a task's physical record and its backlog row as two independently timed steps, so a crash or a forgotten follow-up could leave the two disagreeing: a record with no in-flight row, an in-flight row with no owner, or a finished task still shown in flight. Fold each backlog transition into the script that performs the physical change, under the per-task lock it already holds and before it reports success. Dispatch moves the item to In flight after publishing the task record and fails loudly, removing its provisional record, when that transition cannot land. Completion records an authoritative close and performs it before removing the record, so an interrupted cleanup can be finished later, and its closing message now confirms what already happened rather than instructing a future step. Add a same-home reconciliation sweep to session start so a home that was interrupted mid-transition settles its own books on restart, replaying a recorded close and restoring an in-flight row it already owns a worker for. It never reads or writes another home; the fleet snapshot and the cross-home nudge stay as backstops. Close records are validated before they are trusted: the file is read as raw bytes and rejected outright when it carries a NUL or other control byte, every field must be well formed and non-duplicated, the id must match the record it was found under, the data location must resolve inside this home, and each close argument must carry a permitted, well-formed value. Writer and reader share one validator so a record this home publishes always remains replayable, independent of locale. Homes configured for a manual backlog, and homes with no backlog at all, stay exempt and are unaffected. * no-mistakes(review): Remove stale bootstrap migration helper invocation * no-mistakes(review): Preserve pending closes and narrow signal deferral * no-mistakes(review): Record close before destructive teardown * no-mistakes(review): Refuse pending closes before creating resources * no-mistakes(review): Guard relaunches and preserve cleanup warnings * no-mistakes(review): Reject symlinked records and clarify cleanup guidance * no-mistakes(review): Align dispatch eligibility and protect close replay * no-mistakes(review): Unify exact task incarnation parsing * no-mistakes(review): Render resolved configured backlog path * no-mistakes(review): Harden transition path boundaries against symlinks * no-mistakes(review): Validate lifecycle state before resource actions * no-mistakes(review): Enforce transition tooling and continuous state locks * no-mistakes(review): Consolidate same-home lifecycle file boundaries * no-mistakes(review): Enforce canonical lifecycle containment and tooling contracts * no-mistakes(review): Reject final-component lifecycle record symlinks * no-mistakes(document): Document lifecycle record path boundaries * no-mistakes(lint): Quote literal done tokens in atomicity tests * no-mistakes(ci): Fixed all PR-caused CI failures: bootstrap now treats an absent state directory as an empty fresh home while retaining unsafe-state checks; nested remote secondmate retirement accepts records already removed with the retired home; teardown fixtures now provide valid data/manual-backend configuration; and the manual reminder assertion checks the configured absolute backlog path. Verified the reported tests, remote lifecycle E2E, backlog atomicity suite, Bash syntax, diff checks, and ShellCheck. The documented pre-existing captain-hold failure was intentionally untouched * no-mistakes(ci): Fixed Behavior portable serial 3 by adding `od` to the teardown test’s lsof-free PATH fixture. The new close-record validator legitimately requires `od`; its omission caused teardown to fail before process-group cleanup and stall the shard. Verified the full `tests/fm-teardown.test.sh` suite passes, plus Bash syntax, ShellCheck, and `git diff --check` * no-mistakes(ci): Fixed close replay to durably retain incomplete-cleanup evidence before removing task metadata. Subsequent retries now emit the reconciliation warning even after a backlog probe or close failure. Updated the behavioral regression and verified the full atomicity suite under stock macOS Bash 3.2, plus shellcheck and diff checks * fix(records): validate record bytes without an uncurated tool The byte validation added for close records and directory paths shelled out to od. The spawn and teardown lifecycle runs under a curated command set that deliberately excludes it, so on any restricted PATH the check could not run, the data directory read as unresolvable, and dispatch and cleanup refused - wedging the lifecycle rather than protecting it. An earlier attempt made the failing test pass by adding od to that curated set. That fixed the test to agree with the defect and quietly widened the contract the fixture exists to pin, so it is reverted here. Inspect the bytes with perl instead, which is already in the curated set and already used in this repo for the same portability reason. The emitted values are identical to od's, so the rejection semantics are unchanged: NUL and other control bytes are still refused, legitimate paths containing spaces or non-ASCII characters still round-trip, and the check stays independent of the process locale. The restricted-PATH teardown case now passes because the validator no longer needs od, not because the fixture was loosened. * no-mistakes(review): Enforce dispatch eligibility and atomic remote record publication * no-mistakes(document): Document dispatch eligibility and cleanup alerts
…3342) * fix: publish promote and Relay meta rewrites through contained replace Bare mv still rewrote live task records in place, so a symlink meta could be followed to a target outside state/. Route those field rewrites through the shared publisher and drop the unused library aliases. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(review): Refuse dangling symlinks during X metadata clear * no-mistakes(review): Refuse unsafe metadata before follow-up and promotion side effects * no-mistakes(review): Exercise dangling symlink refusal through clear helper --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…d#2877) * fix(watch): absorb a turn-end whose pane churned since the previous poll The watcher's "absorb a benign turn-end when the crew is provably working" triage was structurally unreachable for any harness whose semantic busy state has no verified source. crew_absorb_class only reports working for an actively running no-mistakes step or an exact busy verdict, and bin/fm-crew-state.sh can only answer unknown for such an adapter, so codex crewmates surfaced a signal wake at every turn boundary with nothing to act on - a full supervisor drain, inspect and acknowledge turn per worker turn, scaling with the number of workers in flight and drowning the wakes that matter in identical noise. Widen the proof rather than bound the wake rate. A wake carrying only bare turn-ended markers is now also benign when the task's pane content changed since the previous poll, compared against the same state/.hash-* marker the staleness backbone already records and already trusts as liveness. That evidence claims no harness semantics, so it fabricates no busy verdict an adapter has not earned, and it needs no adapter cooperation. Absorb stays evidence-driven in both directions. A wake naming any status file keeps the strict proof, every captain-relevant verb still surfaces immediately, and an unresolvable task, a missing prior hash, a failed or empty capture, or an unchanged pane all surface exactly as before. The absorb defers rather than swallows: a crew that has stopped renders nothing further, so its now-static pane surfaces through the staleness backbone within a poll or two. Bounding the surfacing rate instead would have suppressed genuinely stopped workers. The derivation lives with the .hash-* marker format in bin/fm-watch.sh, which owns it, and costs one bounded capture reached only for a no-verb turn-end whose crew is not already provably working. * no-mistakes(review): Captain, guard pane-churn absorption from collisions and secondmates * no-mistakes(review): Captain, make watcher marker identities injective * no-mistakes(review): Captain, isolate ambiguous legacy markers and restore Herdr sourcing * no-mistakes(review): Captain, localize pane-churn collision guard * no-mistakes(review): Captain, reject malformed pane-churn hashes * no-mistakes(document): Document pane-churn turn-end evidence * no-mistakes: apply CI fixes * fix(watch): gate and bound the pane-churn turn-end absorb Make the pane-churn form of positive work evidence opt-in per home and bound how long it may defer one endpoint's bare turn-ends. Absorbing a bare turn-end on pane churn is now reached only when the home creates config/turnend-churn-absorb. The other two proofs read a verdict the harness itself vouches for, while this one infers execution from rendered bytes, so widening the absorb is a home's choice rather than a default every fleet inherits. With the flag absent the predicate returns on its first line and triage is unchanged. Churn and pane staleness read the same pane, so neither can be the other's only backstop. A pane that renders continuously never presents the two consecutive identical hashes the staleness backbone needs, so an unbounded churn absorb left a worker that had genuinely stopped behind such a renderer with no path to surface at all. One endpoint's turn-ends may now ride churn evidence for at most FM_TURNEND_CHURN_ABSORB_SECS, tracked in state/.churn-since-*, after which the wake surfaces and the window restarts. The bound is evaluated before any .stale- state is touched, so a wake that surfaces there leaves the staleness backbone's own classification alone. Covers both with behavioral tests: the same churning fixture that absorbs with the flag surfaces and queues without it, and a spent deferral window surfaces and restarts. The four existing safety guards now run with the flag enabled so they keep proving their specific guard. * no-mistakes(review): Fail closed on invalid churn deferral state * no-mistakes(review): Validate persisted churn deadlines before arithmetic * no-mistakes(review): Make churn deadlines transactional and bounds safe * no-mistakes(review): Compose turn-end evidence per task from one snapshot * no-mistakes(review): Restore strict turn-end fallback guards * no-mistakes(document): Clarify pane-churn supervision documentation * no-mistakes(lint): Fix watcher arithmetic lint issues * no-mistakes: apply CI fixes * no-mistakes(document): Clarify pane-churn fail-closed documentation * fix(bin): prioritize active pipeline-owned crew runs (kunchenguid#3194) * fix(bin): bind the live pipeline-owned run instead of a superseded failed row fm-crew-state.sh bound a superseded FAILED no-mistakes run to a task instead of the LIVE replacement run: the live run's pipeline-owned lane head is not a git object in the task worktree, so head-equality attribution rejected it and the coarse runs-list fallback silently continued past the RUNNING row onto an older failed row whose head equalled the stale worktree HEAD. The home summary then flipped invalid and Bearings hid the home's live work (F10). Attribution precedence now follows the daemon's own identity: - An ACTIVE run for the task's branch binds without head equality while branch_sync.state is pipeline_owned (fm_nm_run_is_pipeline_owned_active); the pipeline owning the branch is itself the attribution. - A genuinely failed run with no later run on the branch still reports failed through the unchanged head-equality path - real failures are not hidden. - In the coarse runs scan, an unresolvable head is unknown attribution and stops the scan (fm_nm_head_resolvable) instead of falling through to an older row; a resolvable-but-mismatched head keeps the historical reused-branch skip. The exemption never applies to a terminal run and requires pipeline_owned specifically, both pinned by negative-control tests. Fixture shape verified against the live incident run's real axi status output. * no-mistakes(document): Updated run-attribution documentation ownership * no-mistakes(review): Captain, make watcher marker identities injective * no-mistakes(review): Captain, localize pane-churn collision guard * no-mistakes(review): Compose turn-end evidence per task from one snapshot * no-mistakes(review): Restore strict turn-end fallback guards * no-mistakes(document): Align pane-churn watcher documentation * no-mistakes(ci): Captain, fixed the flaky cooldown boundary test by freezing its executable clock. The failure reproduced before the fix and passed five consecutive full-suite runs afterward. Extended ShellCheck passed; full lint stopped because actionlint 1.7.12 is not installed --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com>
* fix(bin): add a safe owner for custom-check retirement Agents were improvising rm of check files with unset STATE/ID, which wedges headless panes. Unregister validates the id and state directory first. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(review): Refuse explicitly empty custom-check state overrides * no-mistakes(document): Document custom-check retirement safety contract --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…o dedicated scripts (kunchenguid#3221) * Add quota exhaustion detection and safe fallback helpers - bin/fm-procevent-quota.sh: generic procevent adapter that arms a recurring quota-axi --json poll and wakes firstmate when a tracked provider's effectivePercentRemaining drops below a threshold or its runway.status becomes exhausted_now. - bin/fm-quota-choose.sh: worker-side helper that picks the first ranked harness:model candidate with positive effectivePercentRemaining. - AGENTS.md and .agents/skills/quota-array-dispatch/SKILL.md: document the new helpers and the mid-task quota-exhaustion wake path. - tests/fm-quota-choose.test.sh: unit tests with a mocked quota-axi JSON source. * no-mistakes(review): Fix quota polling and scope bounds * no-mistakes(review): Enforce safe default quota selection * no-mistakes(review): Handle decimal quota values safely * no-mistakes(review): Fail closed on invalid quota inputs * no-mistakes(review): Reject empty quota candidate segments * no-mistakes(review): Harden quota parsing and timeout ownership * no-mistakes(review): Reuse captured quota snapshots consistently * no-mistakes(review): Match quota using explicit candidate providers * no-mistakes(review): Centralize fail-closed quota schema validation * no-mistakes(review): Reject out-of-range quota percentages * no-mistakes(review): Validate quota runway status enum * no-mistakes(review): Tighten quota scope and status contracts * no-mistakes(review): Preserve unknown quota and exact product bounds * no-mistakes(review): Preserve provider-level unknown quota * no-mistakes(review): Reuse canonical verified harness validation * no-mistakes(document): Document mid-task quota handling * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * fix(docs): restore default routing contract, keep quota helper optional Restore the AGENTS.md section 4 always-loaded routing paragraph the PR had deleted, so the standing TOON-first intake, spendPriority ranker, every-candidate accounting, and load-trigger contract stay exactly as before this PR. The mid-task quota wake is optional and must not alter default routing. Restore the quota-array-dispatch skill ownership line to section 4 as the always-loaded intake boundary owner; keep the worker-side helper section as an addition only, without rewiring ownership or load triggers to section 13. * fix(bin): use harness-keyed quota matching in optional helper Revert fm-quota-choose.sh from harness:provider:model tuples back to harness:model candidates with harness-keyed provider matching, per the resolved ask-user finding. The helper is optional; authoritative multi-provider routing (provider discovery from the harness catalog and quota matching by that explicit provider) stays owned by AGENTS.md section 4 and the quota-array-dispatch skill intake procedure, not the helper. Document the multi-provider limitation in the helper header and the quota-array-dispatch skill: the helper maps each harness to one primary provider family only, so a candidate whose established provider differs from that primary family is checked against the wrong quota row. Use it only when the brief fixed the candidate order and every candidate's provider is the harness's primary family. The helper still consumes one already-captured default-TOON or JSON snapshot via stdin or --snapshot and never calls quota-axi itself, so it selects from the same quota state as the intake. * no-mistakes(review): Fix Muse quota mapping and helper contract docs * no-mistakes(review): Reject known-empty quotas and map quota tests explicitly * no-mistakes(review): Preserve unmeasured candidates and enforce snapshot reuse * no-mistakes(review): Fix quota retirement and dependent regression coverage * no-mistakes(review): Accept zero-row quota TOON snapshots * no-mistakes(review): Enforce quota semantics status consistency * no-mistakes(review): Veto dispatch on any exhausted applicable scope * no-mistakes(review): Record exhausted quota scope in wake details * no-mistakes(review): Fix quota help and control dependency coverage * no-mistakes(review): Decode quoted TOON fields and document quota wakes * no-mistakes(review): Validate zero-row TOON and map timeout coverage * no-mistakes(review): Reject multi-value JSON and malformed TOON envelopes * no-mistakes(review): Validate complete nonzero TOON envelopes * no-mistakes(review): Accept producer-shaped quota TOON envelopes * no-mistakes(review): Support empty quota arrays and validate counted rows * no-mistakes(review): Harden TOON completion, scopes, and quoted fields * no-mistakes(review): Preserve unknown-headroom exhaustion and reject trailing fields * no-mistakes(review): Allow unknown headroom under known semantics * no-mistakes(review): Reject noncanonical quota identities * no-mistakes(review): Preserve empty quota polling and validate attention identities * no-mistakes(review): Reject noncanonical provider watches * no-mistakes(review): Validate all candidates before quota selection * no-mistakes(document): Correct quota helper safety documentation * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes
* fix(bin): keep typed Lavish comments when an element is also annotated read preferred element text over prompt, so an annotate-and-comment item dropped the captain's words. Surface prompt as its own field. Co-authored-by: Cursor <cursoragent@cursor.com> * no-mistakes(review): Filter non-comment prompts from Lavish reader output * no-mistakes(document): Clarify Lavish comment presentation contract * no-mistakes(ci): Fixed Lavish reader comment provenance: non-choice prompts are now emitted even when identical to element text. Added observable regression coverage for identical selector+comment input while retaining pure annotation/message coverage. Reader cases, bash syntax, and diff checks pass. Full fm-procevent suite stops earlier at unrelated “reconcile never claimed” setup failure * no-mistakes(ci): Fixed duplicate pure-annotation prompts by emitting `prompt:` only when it differs from captured element text. Updated behavioral coverage for selector+comment, pure annotation, and pure message cases. Focused reader regressions, syntax checks, and diff checks pass. Full suite remains blocked by the pre-existing “reconcile never claimed the registered source” failure * fix(bin): always emit Lavish comments and use real annotation fixtures Stop inferring comment provenance from prompt==text. Real pure annotations have no prompt, so always-emit does not duplicate. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…uid#3420) * Fix public-followup register crashing on empty lock arrays under bash 3.2. bash 3.2 with set -u treats "${arr[@]}" on an empty array as unbound, so the first register in a fresh home aborted before taking the registry lock. The empty-lock regression also runs under the existing stock macOS Bash CI lane so pre-fix code would fail there. * no-mistakes(document): Document stock Bash registration coverage * no-mistakes(ci): Pinned the stock macOS Bash CI lane to tasks-axi@0.2.5, eliminating dependency drift. Verified workflow YAML parsing, git diff checks, and the focused regression under /bin/bash 3.2.57 with tasks-axi 0.2.5 * no-mistakes(ci): Fixed the flaky portable CI test: it treated exited zombie processes as live because `kill -0` succeeds for zombies. The watcher and descendant assertions now check process state and regard zombies as exited. Verified `tests/fm-pr-check-security.test.sh`, ShellCheck, `git diff --check`, and the focused Bash public-followup regression
* fix(herdr): isolate server launch environment * no-mistakes(review): Clear inherited supervision model from Herdr launches * no-mistakes(document): Document Herdr server launch environment isolation
* fix: surface inbound Relay attachments to the responding agent A Discord support thread's screenshots were never seen by the agent handling the mention. The relay delivered them and the poll stashed them: the reporter's images arrived on the `thread_starter` entry of `in_reply_to_chain` while the mention's own media list was empty. The gap was in the responder's playbook, which enumerated a fixed field list (`request_id`, `text`, `in_reply_to`, `in_reply_to_chain`) and so made every other field, attachments included, invisible. Fix it where the gap is, in prose: - Read the complete payload object rather than a fixed field list, so media and later relay fields are never skipped again. - Fetch and view attached media with the agent's own tools, on the mention and on every chain entry, and call out the common shape where only the thread starter carries the screenshots. - Restrict those fetches to known-good platform media hosts over https (Discord: cdn.discordapp.com, media.discordapp.net, images-ext-1.discordapp.net, images-ext-2.discordapp.net; X: pbs.twimg.com, video.twimg.com), report a blocked host instead of working around it, and treat everything fetched as untrusted public input on the same terms as the surrounding thread text. The poll stays out of it and downloads nothing, so no third-party bytes are pulled on the polling path. The new test pins the contract the playbook depends on: a mention in the incident's shape, with an empty top-level media list and screenshots on the thread starter, must reach the inbox with the payload intact and its media URLs unfetched. * no-mistakes(review): Preserve media authority and enforce poll-only fetching * no-mistakes(document): Clarify Relay attachment safety prose
) * Defer inactive startup reconciliation * no-mistakes(review): Queue deferred inactive reconciliation diagnostics durably * no-mistakes(review): Require worker phases to cover startup requests * no-mistakes(review): Make diagnostic wakes safely acknowledgeable * no-mistakes(document): Document deferred startup phase coverage
* fix: bound status presentation lock waits * no-mistakes(review): Distinguish malformed presentation locks from live contention * no-mistakes(review): Bound no-ack drain queue lock acquisition * no-mistakes(document): Document bounded presentation-lock drain behavior * no-mistakes(lint): Annotate bounded lock output global * no-mistakes(ci): Added deterministic regression coverage for successful bounded-lock acquisition after live contention, verifying helper-to-caller PID ownership handoff and caller release. Verified with bash syntax checks, git diff checks, and the full fm-wake-queue test suite
* fix(relay): close a public loop whose work lives in a remote secondmate home A public-followup loop bound to a REMOTE secondmate could never be closed. `clear_public_followup_link` (bin/fm-public-followup.sh:701) required an absolute recorded `work_home_path` for a `secondmate:*` work home, but a remote route has no local path on this machine, so registration records that field empty (bin/fm-public-followup.sh:291). Every close ran that clear first, so `retire` died with "could not clear the legacy X link ... retained for reconciliation" forever, and `deliver` posted the public reply and then stranded the loop at `posted`. `--force` never covered that step. The clear now goes to the remote home over that route's SSH transport, running `fm-x-followup.sh --clear <work-id>` through `bin/fm-on.sh`. The route is decided from `data/secondmates.md` before any local path is consulted, so a same-named local directory can never stand in for a remote home, and registrations already on disk retire without needing a new field. `fm-on.sh` passes ssh's status through, so 255 stays the established "delivered but completion unknown" result this codebase already reconciles: the close is refused, the registration and the remote link are left exactly as they were, and the message names the unknown completion instead of claiming a definite failure. Local secondmate and `main` work homes are untouched, and `--force` still governs only the unresolved-obligation refusal. Three regression cases drive a remote route end to end, faking only the ssh binary at the FM_SSH_BIN seam and then running the real remote entrypoint against a local checkout, so the clear that must reach the remote home actually happens there. * no-mistakes(review): Guard remote link clears by request identity * no-mistakes(review): Fail guarded clears on unreadable remote state * no-mistakes(review): Reject guarded clears on non-writable remote state * no-mistakes(review): Allow no-link retirement in non-writable remote state * no-mistakes(document): Correct public-followup verification guarantee count * no-mistakes(ci): Fixed the guarded link-clear race by ensuring absence is decided under the metadata lock whenever publication is possible. Added a behavioral concurrency regression test. Verified with fm-x-mode and fm-public-followup suites, Bash syntax checks, diff checks, and bin/fm-lint.sh * no-mistakes(ci): Fixed the guarded link-clear race by refusing an unlocked absence decision when a publisher already owns the metadata lock in a non-writable directory. Added a behavioral concurrency regression test. Verified with fm-x-mode, fm-public-followup, syntax/diff checks, and fm-lint * no-mistakes(ci): Fixed the guarded-clear race by refusing all guarded clears when the metadata parent is non-writable, including apparent link absence. Added a behavioral regression with a publisher waiting to create the lock, updated remote-retirement expectations and verification docs. Passed fm-x-mode, fm-public-followup, fm-lint, documentation audience, Bash syntax, and diff checks * fix(relay): bound the guarded remote link clear so it refuses instead of hanging The guarded clear checks that the remote state directory is writable before taking the metadata lock, but that check cannot close the window: the parent can turn non-writable between the check and lock creation, and a lock held by a live holder is indistinguishable from that at the acquire. `fm_lock_acquire_wait` is an unbounded `while ! try; do sleep 0.1; done`, so either case retried forever and `deliver` or `retire` wedged with nothing reported, instead of returning the retained-for-reconciliation refusal the guard exists to produce. This path runs unattended over the secondmate transport, where a wedge is worse than either outcome the guard defines. The guarded clear now acquires through `fm_lock_acquire_wait_bounded` (FMX_LINK_CLEAR_LOCK_TIMEOUT, default 10 seconds) and refuses on timeout through the existing failure path. Unguarded local callers keep the ordinary unbounded wait, so local behavior is unchanged. The bounded primitive's header no longer claims presentation-only scope, since this is a second authorized caller; nothing else in the shared lock infrastructure changed. The regression holds the metadata lock with a genuinely live process while leaving the state directory writable, so the refusal can only come from the bound and never from the writability precondition. Against the unbounded wait it does not terminate at all; with the bound it refuses, retains the registration, writes no receipt, and leaves the remote link untouched. * no-mistakes(review): Harden lock-timeout regression with independent deadline * no-mistakes(review): Restore no-op guarded clears on read-only state * no-mistakes(document): Clarify remote public-followup cleanup contract
) * fix(bin): resolve process-event state roots before validating them The process-event module validated the caller's spelling of a home's state root instead of the directory it operates on: it required the supplied path to equal its own lexical normalization, which rejects any path reached through a symlinked ancestor. On macOS both /tmp and $TMPDIR are symlinks, so an operator home under either could never claim a source. Reconcile still reported the runner started, while the detached runner died writing "cannot claim source" to the discarded stderr, and the source silently never fired. Resolve the state root to its physical directory once, then apply the existing private-directory validation to that resolved directory and derive every path, recorded claim identity, and later confinement check from it. This keeps the confinement contract for the directory actually operated on rather than only for callers that already spelled it physically, and removes the window where an ancestor symlink could be repointed between check and use. Homes already spelled physically behave identically. This was the single cause of both deterministic macOS failures in tests/fm-procevent.test.sh ("reconcile never claimed the registered source") and tests/fm-procevent-when.test.sh ("the winning concurrent arm did not produce an outcome"). The new case pins the behavior with an explicit symlinked-ancestor home, so it fails without the fix on any platform rather than only where the temp root happens to be a symlink. * fix(bin): pin the external capture staging boundary to its physical path The extension capture path pinned its registry staging boundary by comparing `pwd -P` against the caller-spelled registry directory, so a home reached through a symlinked ancestor still refused to start an extension-backed source after the state root itself resolved correctly. That left such a home half working: built-in sources ran while external ones failed. The staging preparer now prints the physical registry directory it validated, matching the inbox and reservation preparers beside it, and the start path pins on that returned path. The new end-to-end case drives the shipped file-signal package from a symlinked home spelling. * no-mistakes(review): Propagate canonical process-event state roots * no-mistakes(review): Propagate canonical state to process-event adapters * no-mistakes(document): Document physical process-event state roots
…kunchenguid#3312) * fix(pi): persist captain outcomes visibly * no-mistakes(review): Recover captain outcomes after cold-start lock acquisition * no-mistakes(document): Document cold-start captain-outcome recovery * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes: apply CI fixes * no-mistakes(review): Prove immediate Pi captain-outcome transcript delivery * 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 * fix(pi): process captain outcomes through a sequence-keyed turn PR kunchenguid#3312 made every captain-facing supervision outcome a durable, exact-once visible transcript entry with the read cursor advancing only after that entry exists. That is the display half of the delivery contract. Left alone it turns a probabilistic silent loss into a deterministic one: the captain sees an anchor line, and firstmate never acts, because nothing opens a turn and nothing records whether main ever processed the outcome. The 2026-08-31 timeline showed the two shapes this must survive on the previous hidden-turn path: seven delivered decision outcomes each answered by an empty assistant message (cursor advanced, no retry, unanswered for close to three hours), and two answered by an unrelated prior reply. Both happened because delivery advanced the cursor at enqueue and accepted whatever the next assistant message was. Add the processing half on top of the persistence half: - bin/fm-branch-outcome.sh keeps a processed marker separate from the read cursor (`unprocessed`, `mark-processed --through`, `processed-init`). It only advances through an explicit sequence-bound acknowledgement, never past the read cursor and never backwards; an absent marker reads as zero and `processed-init` migrates delivered history once so an upgraded home is not re-presented its past. - After the visible entry for a captain outcome exists, the extension hands every still-unprocessed captain row to main as one hidden, typed `fm-branch-process` request listing each `[seq N] task: summary`, opening exactly one main turn. Main closes it only by calling the new `fm_branch_processed` tool with the highest sequence listed. An unrelated, empty, or paraphrased answer leaves the sequence open, and the same request is presented again at the end of the next main run and at session start. The first two presentations of a sequence set open a turn of their own; after that the request rides the captain's next prompt so an ignored request cannot loop, and a session replacement resets that budget. Routine outcomes stay turn-free. - The regressions cover exactly those incident shapes against the real store scripts: an empty answer and an unrelated prior answer neither advance the marker nor stop re-presentation, the acknowledgement is refused beyond the read cursor and outside lock ownership, a partial acknowledgement keeps the newer sequence open, and kunchenguid#3312's own assertions now forbid an unkeyed turn rather than any turn. The store suite pins the marker's bounds and the migration; the real-SDK guard for appendEntry persistence and model exclusion is unchanged. Docs move the protocol from "no model turn" to "one sequence-keyed processing turn closed only by its acknowledgement", and the verification record carries the dated run against Pi 0.84.4. * no-mistakes(review): Harden outcome listing and sequence-bound acknowledgements * no-mistakes(review): Harden outcome state validation and request pacing * no-mistakes(review): Reject unsafe sidecars and unterminated outcome stores * no-mistakes(review): Validate canonical mark-read cursor state * no-mistakes(review): Guard cursor advancement against corrupt processed state * no-mistakes(review): Bind acknowledgements to active processing requests * no-mistakes(review): Reset pacing when processing sequence membership changes * no-mistakes(review): Enforce silent outcome invariants at storage boundary * no-mistakes(document): Document hardened captain outcome processing contracts --------- Co-authored-by: kunchenguid <kun@kunchenguid.com>
…3481) * feat: bound Bearings remote ledger collection * no-mistakes(review): Clarify default remote-ledger collection behavior * no-mistakes(review): Detach reconcile delivery from watcher loop * no-mistakes(review): Enforce bounded snapshot and request captures * no-mistakes(review): Bound legacy summary capture before parsing * no-mistakes(review): Bound primary remote ledger captures * no-mistakes(document): Correct snapshot and reconcile documentation * no-mistakes(lint): Fix ShellCheck quoting in bounded collector * no-mistakes(ci): Fixed all three CI failures: updated the macOS Bearings assertion to 44 tests, made the home-summary test deterministic and aligned with default ledger consumption, and increased the asynchronous reconcile retirement wait for loaded CI. Verified both focused suites, all 44 Bearings tests, ShellCheck, actionlint, Bash parsing, and git diff checks * test: await reconcile request retirement * no-mistakes(review): Avoid empty reconcile queue process churn * no-mistakes(review): Read ledger summaries from immutable snapshots * no-mistakes(review): Reject multi-document home ledger streams * no-mistakes(review): Coalesce durable reconcile requests per target * no-mistakes(review): Unify reconcile keys and reject snapshot streams * no-mistakes(review): Key reconcile requests by stable target ID * no-mistakes(document): Document per-target reconcile request coalescing * no-mistakes(lint): Remove unused snapshot summary file variable * no-mistakes(ci): Adjusted the concurrent collector regression’s end-to-end timing ceiling to account for stock macOS process/jq overhead outside the three-second remote collection budget, while remaining below the 15-second serial-read floor. Verified with stock /bin/bash 3.2: all 44 Bearings tests pass; bash syntax and git diff checks pass * no-mistakes(ci): Fixed legacy summary validation to require exactly one top-level JSON document and added behavioral regression coverage. Stabilized CI by conditionally waiting longer for durable reconcile delivery and synchronously stopping the fm-on worker tree before fixture cleanup. Removed a redundant flaky healthy-path timing assertion; the wedged-reader test still proves concurrent bounded collection. Verified fm-bearings-snapshot, fm-secondmate-reconcile, and fm-on tests, plus project ShellCheck, bash syntax, and git diff checks
* fix(ci): rebalance the portable serial shards on measured durations The "Behavior portable serial 3" shard ran 17-20 minutes against its 20-minute job cap and intermittently timed out seconds after a passing test, on branches and on main alike. Shards are packed longest-processing-time from per-script duration hints, and those hints were last measured on 2026-08-21 at 116 scripts. The lane has since grown to 139 scripts and from ~42 to ~63 minutes: 17 scripts had no hint at all and fell back to the 20 s default, and several existing hints were low by 2-5x (fm-watch-triage 142 s hinted vs 263 s measured, fm-public-followup 36 s vs 197 s). The partition therefore looked perfectly balanced in hint space, 734.6 s per shard, while really running 11.5, 13.6, 18.8 and 16.5 minutes. Script-count balance, which is what the tests asserted, stayed normal throughout and hid it. Refresh the hints from the timing artifacts of three green runs, taking the slowest measurement of each script so the balance holds on a slow runner, and split the lane across five shards instead of four. Replayed against those runs' real per-script durations the worst shard is now 12.54 minutes, 63% of the unchanged 20-minute cap, and the serial lane's wall clock drops from ~20 to ~12.5 minutes. Bound the drift that caused this rather than relying on the hints being refreshed by hand: the coverage guard now reports the unmeasured share as serial_unhinted= and refuses past PORTABLE_SERIAL_MAX_UNHINTED_PERCENT, which leaves room for newly added tests while making a stale table fail the guard instead of silently pushing one shard into its cap. No test changes what it asserts and no test stops running; only the partition across shards changes. * no-mistakes(document): Clarify conservative shard timing aggregate
…uid#3491) * fix(pi): fall back after settled branch errors * no-mistakes(review): Detect provider errors across prompt compaction * no-mistakes(review): Preserve in-flight branch state across selection changes
* fix(pi): recover supervision branch after cooldown * no-mistakes(review): Defer branch recovery until prompt settlement * no-mistakes(document): Clarify supervision cooldown recovery contract
* refactor: remove legacy remote summary reads * no-mistakes(document): Document ledger-only snapshot reads * no-mistakes(ci): Fixed the snapshot test fixture so ledger refreshes use the same fake executable PATH as the snapshot consumer. This preserves observable endpoint freshness after removing legacy summary computation. Verified stock Bash parsing and all 44 Bearings tests pass under /bin/bash; git diff checks pass * no-mistakes(ci): Fixed the CI-only snapshot fixture failure by ensuring the bounded-ledger refresh uses its fake tmux backend. This removes host tmux availability as a source of nondeterminism. Verified all 44 Bearings tests pass, Bash syntax passes, and git diff checks are clean * no-mistakes(ci): Fixed CI nondeterminism in the Bearings fixture: all local ledger refreshes now use the fixture’s fake tmux backend when available, instead of depending on host tmux state. Verified stock /bin/bash syntax, git diff checks, and all 44 Bearings tests with a deliberately failing host tmux
…henguid#3498) * fix(pi): rearm watcher after session replacement * no-mistakes(review): Queue actionable closes across Pi session replacement * no-mistakes(review): Stop replacement arm when handoff persistence fails * no-mistakes(review): Preserve actionable wakes through branch and late child races * no-mistakes(review): Surface late handoff failures without crashing Pi * no-mistakes(review): Coordinate replacement delivery settlement and unique handoff tokens * no-mistakes(review): Retry stale deliveries and release settled claims * no-mistakes(review): Distinguish branch settlement and retry handoff cleanup * no-mistakes(review): Deduplicate persistent handoff cleanup alerts * no-mistakes(review): Acknowledge watcher follow-ups only when consumed * no-mistakes(review): Persist idle follow-ups until agent consumption * no-mistakes(review): Preserve pending outcomes when handoff persistence fails * no-mistakes(review): Arm replacement before awaiting prior delivery settlement * no-mistakes(review): Adopt pending handoffs after lock reclamation * no-mistakes(review): Prevent stale generations from adopting replacement handoffs * no-mistakes(review): Scope replacement handoffs by watcher state * no-mistakes(document): Clarify replacement handoff documentation * no-mistakes(ci): Fixed the failing branch-extension tests to model the new settlement-promise contract. Failure cases now assert that delivery ownership returns to the watcher instead of expecting direct extension fallback. Verified the updated branch suite, Pi watcher suite, shell syntax, and diff checks * no-mistakes(review): Update branch settlement tests and preserve chunked outcomes * no-mistakes(document): Document watcher-owned replacement handoffs * no-mistakes(document): Verify replacement handoff documentation * test(pi): cover watcher-owned branch fallback * no-mistakes(document): Refresh watcher-owned fallback documentation
…d#3495) * fix(bin): resurface terminal statuses lost after branch handling * test(watch): canonicalize process-event fixture homes * no-mistakes(review): Index branch outcomes by causal status position * no-mistakes(review): Recover outcome indexes and deduplicate resurfaced statuses * no-mistakes(review): Handle legacy ambiguity and oversized status diagnostics * no-mistakes(review): Keep unclassifiable oversized statuses silent * no-mistakes(document): Document lost-wake outcome backstop * no-mistakes(document): Update outcome backstop documentation * no-mistakes(ci): Fixed CI regressions in wake-drain: parseable reserved-key decisions can no longer bypass the durable decision-fold guard, and status output is prepared and receipt-committed before presentation to prevent repeated one-shot outcomes after later failures. Added a behavioral regression for receipt commit failure and retry. Targeted backstop, correlation-token, decision-cursor, open-decision, unread-status, syntax, and diff checks pass locally. Shard-4 failures appeared unrelated/flaky; the network-parallel test passed locally * no-mistakes(ci): Fixed the Greptile P1 data-loss issue by committing presentation receipts only after prepared output reaches stdout. Added behavioral coverage proving output failure leaves the backstop retryable and receipt failure may duplicate but never lose a presentation. Relevant wake-drain suites and syntax/diff checks pass. The shard-4 Pi extension failure is unrelated to this PR and did not warrant changes * no-mistakes(ci): Stabilized tests/fm-bootstrap-network-parallel.test.sh by replacing scheduler-sensitive equal-sleep timing with bounded synchronization between mocked fetch and remote probes. This preserves detection of real serialization while avoiding false failures under CI load. Verified with five consecutive test runs, bash syntax validation, ShellCheck, and git diff checks. The separate Pi stock-rendering failure reproduces locally but is unrelated environment/version drift * no-mistakes(ci): Fixed Behavior portable serial 4 by adding fm-classify-lib.sh and fm-timeout-lib.sh to the broken-root Pi test fixture; fm-branch-outcome.sh now depends on them. Verified the full Pi branch-extension suite with real-Pi checks skipped, the wake-drain outcome-backstop suite, Bash syntax, and git diff checks. Greptile findings are already addressed at HEAD; the no-mistakes attestation failure is external head-SHA state
…id#3503) * fix(bin): deliver typed terminal results from remote work homes A public commitment whose work is bound to a REMOTE secondmate home could never receive its typed terminal result. `fm-public-followup.sh brief` printed an emit command carrying this home's own absolute path and this checkout's own script path, neither of which exists on the machine the worker runs on, so the worker had nothing it could write to that the owning home would ever read - and `consume` kept finding nothing while the promise stayed open. The brief is now route-aware: for a remote work home it prints that route's own code root and home with `--stage-in`, so the typed event is staged in the home where the work actually runs, and the closing paragraph names the owning home as the one on the other machine instead of pointing at the path above it. The owning home collects those staged results over the same SSH route it reaches that secondmate on, because the transport only runs outbound: `consume` pulls them into its own inbox and reconciles them exactly as it reconciles a local report. Collection is non-destructive until the result is durably held, so a dropped connection cannot lose a terminal result, and a route that could not be reached is named in `consume`'s output with the promise left open rather than reported as an empty inbox. A local work home is untouched: the brief still prints `--home` with this home and this checkout's script, and the event still lands directly in this home's typed terminal-result inbox. This is the emit-side counterpart of the retire/clear fix in kunchenguid#3479 and reuses the remote-route resolution that landed with it. Reconciling a loop bound to a remote route now reaches that route, so the existing remote cases drive `consume` through the same faked transport their other steps already use. * no-mistakes(review): Fail loudly on unresolved routes and invalid staging homes * no-mistakes(review): Fail collection when remote outbox is unreadable * no-mistakes(review): Surface reassigned remote routes during empty collection * no-mistakes(review): Fail remote collection on invalid registrations * no-mistakes(review): Reject unsafe registration entries during remote collection * no-mistakes(review): Restore healthy empty remote collection behavior * no-mistakes(review): Skip remote collection for delivered registrations * no-mistakes(review): Skip delivered registrations before route validation * no-mistakes(document): Document remote follow-up collection semantics
…#3504) * fix(bin): exclude secondmates from home-summary child inventory kind=secondmate meta records never have backlog rows, so counting them in unowned_children or terminal_in_flight made a clean main home look invalid once earlier ledger checks passed. * no-mistakes(review): Cover terminal secondmate in-flight exclusion * no-mistakes(ci): Updated the stock macOS Bash CI snapshot expectation from 15 to 16 tests. Verified all 16 snapshot/fleet-view tests pass under Bash 3.2.57 and `git diff --check` succeeds
* fix(bin): self-heal status-outcome indexes on every drain Missing ready markers were skipping the lost-wake backstop on non-Pi homes because only the Pi branch ran processed-init. Drain now rebuilds those indexes under the outcome lock and fails closed only on a real store fault. * no-mistakes(review): Guard held-lock initialization and fail marker writes * no-mistakes(document): Document cross-harness outcome-index self-healing
…nchenguid#3505) * fix(bearings): keep active children underway beside a captain hold Project each readable home's active children into Underway independently of the home-level captain-decision classification so a hold no longer hides live work. * no-mistakes(review): Preserve Underway repos and disclose child truncation * no-mistakes(review): Fall back to task project for Underway repos * no-mistakes(ci): Updated the stock macOS Bash CI assertion from 44 to 45 Bearings tests, matching the newly added behavioral regression. Verified all 45 tests pass under /bin/bash, Bash syntax checks pass, and git diff validation is clean
…enguid#3513) * fix(pi): settle watcher delivery on Pi accepting the follow-up A follow-up queued while main is streaming joins the running run without ever raising before_agent_start, so waiting on that event before clearing the successor pipeline (kunchenguid#3498) stalled every later actionable close: no successor started, no wake was delivered or offered to the branch, and the turn-end guard woke main to re-arm by hand after every close. The pipeline now settles once Pi accepts the follow-up. Consumption is observed at before_agent_start for an idle main and at the user message_start for a streaming main, and decides only what a replacement session (/new, /resume, /fork, reload) replays. An exhausted restoration delivers its typed failure without launching an arm past the retry bound, which the stall had hidden. The replacement-coordinator map is typed so the strict no-emit typecheck passes again. Tests: the doubles no longer raise before_agent_start for a streaming send, a portable regression drives two actionable closes while main streams and proves the successor chain plus consumption-scoped replay, and a credential-free real-SDK probe pins Pi's event contract for both the streaming and the idle follow-up. Claude-Session: https://claude.ai/code/session_01QJjTsUvKkWAwLGNoncaZ3a * fix(pi): retry a verified successor that fails during wake delivery A verified successor can exit while the wake it was started for is still being delivered, most plausibly during a branch turn that holds the settlement for minutes. Its failure close arrived while the pipeline's single-flight guard was set, so the close handler skipped the retry, and the pipeline's end no longer launched an arm, which left the live generation with no watcher and no retry timer. The close handler now records that failure when the child had reported readiness and was not retired by the restoration itself, and the pipeline runs the ordinary bounded, lock-checked retry for it once the delivery settles. A restoration started for a later pending supersedes it, and an exhausted restoration still hands repair to main without a further arm. The regression holds a branch settlement open while the verified successor exits with a failure and proves one retry watcher starts after the settlement releases, none while it is held. Claude-Session: https://claude.ai/code/session_01QJjTsUvKkWAwLGNoncaZ3a
claude-code-cli (pi-cli-mate) respawns a fresh, non-persistent CLI session on every ordinary Pi turn and requotes the full prior conversation as text, so it never reuses prompt cache across turns; a live repro confirms real token cost for this pattern rather than a counting bug. xai/grok-4.3 was verified to reuse cache across sequential turns in the same session. Neither traces back to Firstmate's own session-start/compaction machinery, which stays capped at one delivery per generation, 512 KiB max.
Traced why a session-start report showed auto-compact firing at a displayed 4.4% context usage: AgentSession._checkCompaction() falls back to a raw chars/4 re-estimate of the entire message array whenever the just-finished response's own usage sums to zero, which is a different source than the footer's last-known-good-usage percentage. The staleness guard meant to catch this only applies when a prior compaction and a valid usage index both exist. Provider-independent; not exercised end to end here, recorded for follow-up.
…TH lookup claudeBinary() resolves the nested CLI via a bare "claude" name against whatever PATH the enclosing Pi process inherited, with no PI_CLAUDE_BIN override set anywhere in Firstmate. bashrc's standard non-interactive-shell guard skips the ~/.local/bin PATH addition for any bash invocation without an interactive flag. fm-pi-primary-restart-lib.sh already resolves the pi/ pi-signed binary to an absolute path for exactly this reason; claude-code-cli never applies the same pattern to its own nested spawn. Records the connection to the auto-compact trigger-divergence finding: a failed spawn produces the zero-usage response that trips it. Not independently reproduced.
Ran Pi's own exported compaction functions (calculateContextTokens, estimateContextTokens, shouldCompact, getLatestCompactionEntry) directly against a constructed message array matching the reported shape: a real 6.6%-of-window last usage, one action's worth of large trailing content, then a zero-usage error response. The trigger computed 146.6% from the same array the footer reads as 6.6%, and crossed the auto-compact threshold - matching both the 7%-displayed auto-compact report and the 6.6%->150.2% jump report. No value is added to itself; shouldCompact() carries no state between calls, so what can persist across a failed compaction retry is the triggering condition (large trailing content untouched by a failed compact), not a stale counter. Also records that the already-landed ENOENT fix (30c0891) closes one concrete way to produce the triggering zero-usage response, not the Pi-core divergence itself.
A further live report showed auto-compaction firing with no active worker or heavy turn in progress. Grepped agent-session.js for persistent instance state across the compaction paths: the only one, _overflowRecoveryAttempted, scopes only the separate overflow-recovery case and resets per turn. The threshold path this doc already reproduces reads no instance field and shouldCompact() takes only its call arguments, so nothing is cached or made more aggressive by an earlier failure - the divergence depends on that turn's own response usage and on how much untouched content already sits in the message array, not on how much work the triggering turn did. This incident's own transcript was not available to trace further.
… reinstall regression Direct source inspection of the running installed bundle confirmed both merged Case 3 trigger and getContextUsage() display fixes were active for roughly 8.5 hours, then a mid-task pi update reinstalled a stock 0.84.4 build that silently dropped the fork-only patch (confirmed via package.json, settings.json, and the bundle's clamp-expression presence before/after). Also measures the isolated harness/tool-schema baseline, the Pi digest's existing 512 KiB/once-per-generation containment, and this host's live default provider/compaction settings to explain the newly reported 48.6%/287%/360.8% jumps without any further code change. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013XPDkmktpahrVQuhPqXow5
Valentino-Sole
pushed a commit
that referenced
this pull request
Sep 4, 2026
…nguid#3681) * fix(bin): recognize active pipeline fix rounds with unfetched run heads A no-mistakes fix round advances the run head beyond the submitted head, and the pipeline commits in its own checkout, so the task copy never receives the new commit object. fm-crew-state's strict head rule rejected the active row, the coarse runs-list scan skipped it and matched the older failed row at the submitted head, and an active validation read as failed (observed on model-routing-benchmark-hardening: active head ac61c64 vs task copy at fb47636d). fm_nm_runs_status_for_worktree in bin/fm-nm-run-lib.sh now owns runs-ledger attribution: the branch's newest row alone decides, and a newest row whose head cannot resolve locally is recognized only as a provable pipeline-owned continuation - active (running) and anchored by the immediately older row for the same branch having ended at exactly this worktree's HEAD. The reader keeps the axi TOON as full detail for that proven same-branch run. Unanchored, ancestor-anchored, and terminal unresolvable rows stay unattributed, so branch-name coincidence and other tasks' runs never match, and fm_nm_head_matches_worktree keeps its exact prior semantics for teardown (verified by the full teardown suite). Tests: reproduction regression for the unfetched active fix head (reads working via full run-step detail), coarse-path continuation when axi answers another branch, and negative controls for the unanchored active row and the unresolvable terminal row with the historical fallback preserved. Ported onto upstream/main f4d7875, where kunchenguid#3194 independently added the branch_sync custody exemption on the full axi-status path: both mechanisms now coexist, each owning one surface (TOON custody on the full path, the runs ledger on the coarse path). The port deletes the superseded coarse scan-and-skip (nm_runs_status_for_branch) and its now caller-less helpers (fm_nm_head_resolvable, nm_coarse_head_matches_worktree), renames the exemption comment's "the one exemption" phrasing now that a second complementary exemption exists, and points the stale FM_CREW_STATE_RUNS_LIMIT comment at fm_nm_runs_status_for_worktree (judge follow-up #1). The parent coarse-guard test's fixture is the ledger-anchored continuation shape, so its expectation flips to the fixed behavior (working via run-step, never the older failed row); a new mismatched-anchor coarse negative control preserves that guard's original no-anchor protection (pane answers, never the older row). * no-mistakes(document): Clarify pipeline attribution documentation
Valentino-Sole
added a commit
that referenced
this pull request
Sep 4, 2026
* fix: start a fresh supervision branch for every main session (kunchenguid#3600) * fix(pi): start a new supervision branch conversation per main session The supervision branch reopened one recorded conversation forever, so every main session start reloaded the current generated prompt and then weeks of accumulated thread, where a superseded rule could still outweigh today's. The branch conversation is now scoped to one main session: the session generation owns the recorded conversation, so a cold start, /new, /resume, /fork, or a reload always builds a new one, while a rebuild inside one session (a model or effort change) still continues that session's own conversation. The dialog mirror re-anchors with it. Its durable cursor records what the previous branch conversation received, so a /resume or reload - which keeps main's own session file - would otherwise leave the new branch blind to dialog main itself still has. The reset is bounded by the current main session, and the cursor keeps advancing incrementally within it. The durable outcome store and its processed marker are untouched, so unacknowledged captain-facing outcomes still re-present on the new main session. * no-mistakes(document): Document fresh Pi supervision conversations * no-mistakes(ci): Fixed the flaky concurrent inbox failure. Lock acquisition now retries when a competing lock disappears between a failed claim and inspection. Added a behavioral regression covering that race. Verified the full inbox test four times, project lint, and git diff checks * feat: restart second mates after instruction updates (kunchenguid#3614) * feat(update): restart second mates whose instructions changed /updatefirstmate pulled new bytes onto disk and then asked each advanced second mate to re-read them. A running agent holds AGENTS.md and every loaded skill frozen from launch and no verified harness offers a reload, so that steer could not reach a loaded skill at all and left the mate holding two contradictory copies of its own job description. An eligible mate is now restarted instead, in the same home and endpoint, through the existing transactional relaunch. The restart is gated on the mate first writing down the open work it holds only in conversation - the open-record half of /stow, never its memory sweeps - so an unregistered captain call is flushed before the conversation is spent. Anything that leaves the reload unprovable falls back to the old re-read message and is reported as exactly that, never as a clean reload. Remote mates take the same path: fm-remote-secondmate-control.sh gains a relaunch verb whose host-local leg runs that same control plane, since the mate is an ordinary local secondmate from its host's point of view. The primary resolves the profile and passes it explicitly, because config/secondmate-harness is not inherited and the file on that host belongs to a different home. fm-update.sh now splits its advanced live mates into a restart set and a nudge residual, and both sets require a changed instruction surface, which also closes the over-nudge against the session-start sweep. Restart is stricter still: a bin/-only advance reloads itself on the next call, so it never costs a conversation. Colocated tests cover the gating, the persist-then-restart order, the task-subset persist request, each unsafe fallback, the remote hop, and the remote sync's new instruction-surface report. * no-mistakes(review): Fix restart correlation, concurrent waits, and lifecycle reporting * no-mistakes(review): Parallelize relaunches and classify replacement incarnations * no-mistakes(review): Gate restart actions on live agent state * no-mistakes(review): Handle failed restart workers without hanging * no-mistakes(review): Nudge legacy remotes and preserve persist recovery * no-mistakes(review): Document one-time secondmate restart rollout * no-mistakes(review): Honor arrived replies and refresh remote profiles * no-mistakes(review): Revert remote parent profile reconciliation * no-mistakes(review): Reset remote profile defaults and honor published results * no-mistakes(review): Preserve fallback nudges for unverifiable secondmates * no-mistakes(document): Document second-mate restart update flow * no-mistakes(lint): Fix ShellCheck warnings in restart scripts * perf: accelerate local validation with bounded concurrency (kunchenguid#3644) * perf(tests): route gate verification through the bounded concurrent runner Local validation was the pipeline's dominant cost: across 67 recorded no-mistakes agent sessions on this repo, 99.3% of command execution was `bash tests/*.test.sh`, run strictly one script at a time, and 2% of those calls were killed by an agent-guessed timeout and paid for twice. Three changes, each measured: - `.no-mistakes.yaml` pins `commands.test` to `bin/fm-test-run.sh --changed --exclude-family real-herdr-gated`. The runner already owns changed-file selection, bounded concurrency, the refusal of unproven scripts, and a generous automatic per-script bound, so the gate's baseline is neither a serial chain nor a guessed timeout. It stays intent-targeted - the Test step still runs its evidence agent on top - and excludes the live-Herdr family the required Herdr lane owns. - `bin/fm-test-run.sh` gives a plain list of script paths the same bounded automatic scheduler and automatic bound that `--changed` gets. Naming several subjects is how a verification round asks for exactly those scripts. The curated selections are untouched: `--lane` still composes CI shards whose serial lane must stay serial, `--family` is what the required Herdr lane runs, and `--all` stays a deliberate complete regression. - `pr-forge` is admitted to the concurrent-safe family registry on two consecutive clean proofs. `docs/fm-test-isolation-proof.md` records those, and records `secondmate` and `session-bootstrap` as refused with the exact script and reason each failed on, so the refusals are actionable rather than silent. Measured on this host, 0 failures on both sides: verification round, 4 scripts 448s chained -> 231s through the runner (-48%) pr-forge family 409.2s at 1 worker -> 237.9s at 4 (1.72x) watcher-wake-lock family 1311.1s at 1 worker -> 539.3s at 4 (2.43x) A fourth lever was implemented and then removed because the measurement refused it: raising the bounded-wait sample interval from 0.1s to 0.5s made `fm-watch-triage.test.sh` slower, 435s and 440s against 390s and 393s unchanged, back to back. Those sleeps are not overhead added to the clock - they are how a test waits for a subject moving on fm-watch.sh's own one-second cadence - so sampling less often only delays detection. It also broke `fm-watcher-lock.test.sh`, which catches a transient rather than waiting for a settled condition. CONTRIBUTING.md records that result so the experiment is not repeated. * no-mistakes(review): Separate concurrent runs by isolation proof family * no-mistakes(review): Limit automatic timeouts to changed-file validation * no-mistakes(document): Clarify validation concurrency documentation * fix: copy PR URLs from durable records (kunchenguid#3648) * fix: copy PR URLs from records or abstain, never assemble them Supervision reported a plausible but dead PR link three times because its prompt demanded a full https:// URL at a moment when only a PR number was observable, so the model assembled an owner/repository from memory, and the PR check then accepted that URL and wrote it into the task record, after which the model kept defending its own tool-endorsed guess over the worker's real link. Three changes close that chain without any live forge lookup, so private forges are treated exactly like public ones: - bin/fm-branch-prompt.sh no longer mandates a URL. Its new "PR identity: copy or abstain" section requires a URL to be copied verbatim from a durable record (the done: PR <url> status line, pr= metadata, or the backlog note), forbids assembling owner, repository, host, or number from memory, and has the branch report only the identifier it actually holds when no record names the URL yet, leaving the PR check unarmed until the worker's ready line arrives. AGENTS.md section 7 and 9 carry the same copy-or-abstain rule for main in place of the bare full-URL mandate. - Worker briefs (bin/fm-brief.sh, ship and scout rules) require the full https:// URL wherever a PR is mentioned - status line, terminal, or summary - never a bare "PR 108", so the link is in view as early as the number is. - bin/fm-pr-check.sh refuses, offline and before any side effect, a URL that the task's own done lines contradict, printing both spellings; a log naming no URL still records the argument as before. fm_pr_status_ready_urls in bin/fm-pr-lib.sh owns reading those lines. The refusal also reaches bin/fm-pr-merge.sh, so nothing merges under a contradicted URL. Tests cover the offline refusal with zero side effects, the recorded spelling being accepted, markdown-wrapped and punctuated URLs, working lines not counting, the merge wrapper propagation, a self-hosted merge request with no forge call, the prompt carrying the rule, and the brief carrying the worker rule. * no-mistakes(review): Remove stale PR URL enforcement * no-mistakes(ci): Removed backlog notes as an accepted PR identity source. PR URLs may now be copied only from the task’s `done: PR <url>` status or canonical `pr=` metadata; otherwise supervision reports only the known identifier and leaves PR checking unarmed. Updated related guidance/docs and verified with branch-supervision tests, brief tests, ShellCheck, and `git diff --check` * fix(bin): disable Claude feedback drafts for fleet launches (kunchenguid#3661) * fix(bin): disable Claude's feedback-draft flow for fleet-launched agents Scope --settings '{"feedbackDrafts":"off"}' to every Firstmate-launched Claude crewmate and secondmate, so /bug and /feedback never queue or submit a bug report on the captain's behalf. feedbackDrafts is the documented settings key (Claude Code changelog 2.1.247); the per-launch CLI flag never touches the captain's global settings.json. Claude-Session: https://claude.ai/code/session_01XYAXXzr4oZx9NjZb1veeE3 * no-mistakes(review): Prevent managed settings from re-enabling Claude feedback drafts * no-mistakes(document): Fix Claude feedback documentation formatting * fix(bin): layer both feedback-draft controls for defense in depth The prior --settings-only fix can be overridden by a managed Claude settings policy (feedbackDrafts precedence). Keep CLAUDE_CODE_SEND_FEEDBACK=0 alongside --settings '{"feedbackDrafts":"off"}': either control alone disables the SendFeedback tool, so a managed override of one still leaves the other in force. Claude-Session: https://claude.ai/code/session_01XYAXXzr4oZx9NjZb1veeE3 * no-mistakes(document): Document Claude feedback-draft suppression ownership * feat(tests): run three more validation families concurrently (kunchenguid#3662) * perf(tests): admit three more families to concurrent validation The three families that `docs/fm-test-isolation-proof.md` recorded as refused were not refused for concurrency. Each blocker was a test that decided a property by wall clock, or a script filed where it cannot run. Fixing those three things admits all three families and recovers 28.6 minutes of local validation with no assertion removed or weakened. - `tests/fm-backlog-handoff.test.sh` injected its pre-move crash by killing the handoff, sleeping a fixed second, then delegating the move to the real binary. Nothing ever killed the fake, so on a host slow enough for the case's next assertions to take longer than a second, the orphan woke and completed the very move the case requires left undone, and recovery then failed with `Task "pre-move-crash" not found in this backlog`. Watching the two backlogs during the injected crash showed exactly that, the item moving one second after the crash. All four crash injections in the file now go through a new `fm_fake_crash_injector` shim that signals the target and returns only once it is observably gone, and the pre-move fake never delegates the move at all. - `tests/fm-session-start.test.sh` proved the startup digest does not block on a slow current-state read by timing the whole digest against a fixed eight-second sleep, which a loaded host exceeds without the property being violated. It now holds that read open until the case releases it and asserts, the moment the digest returns, that the read has not finished. A digest that waited would wait indefinitely rather than for an interval a slow host can out-run, so the assertion is stronger than the bound it replaces. Its scan budget moves to the maximum, because the old value left two seconds of margin over the fixed sleep and measured the host rather than the deadline that `tests/fm-inactive-reconcile.test.sh` owns. - `fm-backend-herdr-focus-flash-e2e` was filed in the family map's catch-all, which put it in the portable serial lane, where Linux CI gate-skips it: that real-Herdr regression was running nowhere. It moves to `real-herdr-gated` and the required Herdr lane. `fm-claude-stop-autoarm-live-e2e` gate-skips on its opt-in variable and moves to `live-harness-optin`. The 28 remaining ungrouped scripts become an enumerated `standalone` family instead of admitting `unclassified` itself. `unclassified` is the family map's `*)` arm, so admitting it would silently grant concurrency to every test added afterwards, which is exactly the population with no proof. A new test still lands in `unclassified` and stays serial, and `tests/fm-test-run.test.sh` covers that split behaviorally. Each family passes two consecutive four-worker proofs with zero failures. On the production runner, `secondmate` goes 1233.1s to 453.4s, `session-bootstrap` 756.4s to 286.4s, and `standalone` 724.6s to 261.1s: 2.71x overall and 1713.2s recovered. The whole suite runs 177 scripts in 52.6 minutes of wall clock against 121 minutes of summed script time. * no-mistakes(document): Refresh concurrent validation and shard documentation * no-mistakes(ci): Fixed the real-Herdr focus-flash E2E race exposed by reclassification. Part C now starts its persistent child atomically via `pane run` and verifies stable child identity through Herdr’s public `process-info` interface, avoiding the racy send-text/send-keys sequence and platform-specific `ps` matching. Verified with bash syntax checking, ShellCheck, git diff checks, and the complete E2E test on Herdr 0.8.2 * feat: structure no-mistakes ask-user escalations (kunchenguid#3670) * feat(brief): structure no-mistakes ask-user escalation as event + snapshot file Crewmates escalating a no-mistakes ask-user gate now report one status event naming every finding id plus a snapshot file holding the gate's axi finding records verbatim (id, severity, file, line, description, authority), using the same shape even for a single finding. The status line never paraphrases. The format is defined once in fm-dod-lib.sh and rendered into both the scout and ship rule 6 in fm-brief.sh, so a promoted scout - whose rule 6 fm-promote.sh preserves unchanged - gets the identical contract as a freshly-spawned no-mistakes ship worker. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PpiWaDerbYavTLPPtEjQei * no-mistakes(review): Preserve ask-user escalation output contract * no-mistakes(review): Align escalation format test expectation * no-mistakes(review): Scope ask-user escalation instructions correctly * no-mistakes(review): Remove ask-user from generic decision rules --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> * fix(bin): require self-sufficient no-mistakes intent (kunchenguid#3671) * fix(bin): require a self-sufficient no-mistakes intent A no-mistakes worker's --intent is only as useful as the string it passes. PR kunchenguid#3604 shipped with an intent that was only "do 1, 2, 3, 7 from the report": the real contract lived in a private scout report and never reached --intent, so nobody holding that string plus the codebase could have derived the specification. This is pure instruction at the contract's one owner; no spawn-side or promotion-side check is added. - bin/fm-dod-lib.sh: the generated no-mistakes Definition of done now states that the --intent string must be self-sufficient (the string plus the codebase reconstructs roughly the same specification) and tells the worker to write the substance of any report, decision, or PR the captain's intent refers to into --intent rather than the pointer, while Firstmate build instructions and the worker's own decisions still stay out. The spawn-time overlay points back at that rule so its "supersedes" wording cannot cancel it, and the header's owner statement carries the rule. - AGENTS.md section 11 and bin/fm-brief.sh's header ask Firstmate to include the substance of referenced material when filling ## Captain's intent, and section 11 points at the owner of the rule. - tests/fm-brief.test.sh and tests/fm-task-delivery.test.sh assert the rendered brief and launch contract carry the rule. Claude-Session: https://claude.ai/code/session_01YMhEe42q7BAAoN6RxNuzim * no-mistakes(document): Replace incident-specific intent test commentary * fix: accelerate local Bearings snapshot composition (kunchenguid#3499) * Speed local fleet snapshot composition * no-mistakes(review): Stabilize task inventory during concurrent snapshot composition * no-mistakes(document): Document local snapshot observation concurrency * no-mistakes(ci): Fixed CI failures by making empty task manifests compatible with stock macOS Bash 3.2, snapshotting task metadata before concurrent observations to prevent generation drift, strengthening the behavioral race regression, and updating the stock-Bash Bearings test count to 45. Verified fleet snapshot tests (15), Bearings tests (45), workflow lint tests, project lint, Bash 3.2 parsing, and diff checks * no-mistakes(ci): Fixed the Linux CI failure caused by passing large backlog/task JSON through jq command-line arguments, which exceeded the per-argument size limit. Both inventory projections now stream large JSON inputs through stdin. Verified with fm-bearings-snapshot.test.sh (45 tests), fm-fleet-snapshot-view.test.sh (15 tests), Bash syntax, and git diff checks * no-mistakes(ci): Fixed concurrent task teardown during metadata capture: vanished metadata is now omitted while genuine copy failures remain fatal. Added a deterministic public Bearings regression test and updated CI’s expected test count. Verified with the full Bearings suite, workflow-lint suite, Bash syntax checks, and git diff checks * no-mistakes(ci): Fixed PR-caused CI and review issues: streamed large fleet JSON through jq stdin to avoid Linux argument limits, kept crew-state reads bound to captured metadata generations, and strengthened the behavioral race test. Bearings (46 tests), fleet snapshot (15 tests), crew-state, backend, lint, Bash syntax, and diff checks pass locally. Serial shard 5’s unrelated task-inbox segmentation fault appears infrastructural/flaky * no-mistakes(ci): Fixed endpoint-state generation crossing by validating captured spawn_gen before and after local endpoint probes, falling back to exact metadata identity for legacy tasks. Stale probe results now become unknown instead of false unhealthy state. Added a behavioral relaunch-race regression test. Verified the full Bearings snapshot suite, shellcheck, bash syntax, and git diff checks * fix(snapshot): keep live observations generation-coherent * no-mistakes(review): Keep secondmate observations generation-bound without copying reports * no-mistakes(document): Document generation-coherent snapshot observations * test(bearings): measure local read overlap instead of wall-clock budget The large-local-snapshot regression asserted that a whole snapshot composed in under five seconds. That bound measures how loaded the host is, not whether the per-task reads actually overlap, so it failed intermittently on a contended machine: one run in six on a box at load 16-20, landing exactly on the five second boundary. Time a serialized run and a concurrent run of the same workload instead and require the concurrent one to save at least two seconds. Both runs pay the same composition overhead, so the difference isolates the overlap this change delivers. Five one-second reads serialize into five seconds and overlap into about one, and re-serializing the reads collapses the saving to roughly zero, so the assertion still fails loudly if the concurrency regresses. Also bump the pinned Bearings test count to 48, since rebasing onto the current default branch picked up its captain-hold test. * no-mistakes(review): Restore JSON-derived decision flags * no-mistakes(review): Unify status-derived snapshot observations * no-mistakes(ci): Updated the stock macOS Bash CI check’s Bearings test count from 48 to 49. Verified the full Bearings suite passes and emits exactly 49 TAP successes; git diff checks pass * fix: prevent stale supervision wake loops (kunchenguid#3672) * fix(bin): stop the supervision branch's stale-ack and ghost-report loops Clean-slate implementation of the four authorized recommendations from the supervision-ghost-retrigger analysis (items 1, 2, 3, and 7), in their minimal form, superseding PR kunchenguid#3604: - fm_branch_report refuses a task the wake being handled never named. The extension fixes the reportable task set from the eligible rows before each prompt (signal and stale rows resolve to their tasks, a heartbeat allows any task with a live record, fleet is always allowed), so a report typed from memory about a task whose records teardown already removed is never stored or delivered. - An acknowledgement that consumes nothing says "nothing was acknowledged through N" and prints the exact --ack-through / --recovery-generation command for the current presented wake, instead of "re-run the drain", which re-fed the same stale acknowledgement in a loop. - bin/fm-guard.sh no longer tells the branch actor to drain queued wakes while it is handling them; it names the granted rows instead. - Teardown removes state/.<task>.branch-outcome-index for ordinary tasks and descendants; the index rebuild and the append-side index write both skip a task with neither a live record nor a status log, so the branch's report of a teardown it just performed is stored without recreating the index. No new locking, no spawn-generation binding, and no retired-task refusal: the branch can still report the outcome of a task it just tore down, and the teardown test now proves that path end to end. * fix(bin): narrow the branch report scope and guard silence to the minimal form Apply the four review decisions on the clean-slate branch: - A signal or stale prompt may report only the tasks its own rows resolve to; fleet is refused there too. A heartbeat review is not scoped by task at all, so the extension no longer tracks live task records and refuses nothing by task id during a fleet review. - The outcome-index rebuild no longer skips retired tasks; the append-side skip alone keeps a torn-down task's index from being recreated. - bin/fm-guard.sh keeps the queued-wakes warning silent for the branch actor instead of printing a replacement note. * no-mistakes(document): Align supervision docs with scoped wake handling * fix(bin): avoid fleet snapshot argument limits (kunchenguid#3677) * Fix fleet snapshot large JSON transport * no-mistakes(review): Captain: file-back fleet snapshot transport safely * no-mistakes(review): Captain: file-back parent summary aggregation * no-mistakes(ci): Rebased the PR's three commits onto f4d7875 and resolved the fleet snapshot conflict while preserving the base's task-observation lifecycle. Fixed Greptile's valid finding by recursively removing the private mktemp transport directory, so future transport files cannot cause cleanup to fail. Verified with tests/fm-home-summary-refresh.test.sh, bin/fm-lint.sh, git diff --check, and ancestry checks. All passed; the fix remains as an uncommitted worktree change for the outer executor * fix(bin): attribute active runs with unfetched pipeline heads (kunchenguid#3681) * fix(bin): recognize active pipeline fix rounds with unfetched run heads A no-mistakes fix round advances the run head beyond the submitted head, and the pipeline commits in its own checkout, so the task copy never receives the new commit object. fm-crew-state's strict head rule rejected the active row, the coarse runs-list scan skipped it and matched the older failed row at the submitted head, and an active validation read as failed (observed on model-routing-benchmark-hardening: active head ac61c64 vs task copy at fb47636d). fm_nm_runs_status_for_worktree in bin/fm-nm-run-lib.sh now owns runs-ledger attribution: the branch's newest row alone decides, and a newest row whose head cannot resolve locally is recognized only as a provable pipeline-owned continuation - active (running) and anchored by the immediately older row for the same branch having ended at exactly this worktree's HEAD. The reader keeps the axi TOON as full detail for that proven same-branch run. Unanchored, ancestor-anchored, and terminal unresolvable rows stay unattributed, so branch-name coincidence and other tasks' runs never match, and fm_nm_head_matches_worktree keeps its exact prior semantics for teardown (verified by the full teardown suite). Tests: reproduction regression for the unfetched active fix head (reads working via full run-step detail), coarse-path continuation when axi answers another branch, and negative controls for the unanchored active row and the unresolvable terminal row with the historical fallback preserved. Ported onto upstream/main f4d7875, where kunchenguid#3194 independently added the branch_sync custody exemption on the full axi-status path: both mechanisms now coexist, each owning one surface (TOON custody on the full path, the runs ledger on the coarse path). The port deletes the superseded coarse scan-and-skip (nm_runs_status_for_branch) and its now caller-less helpers (fm_nm_head_resolvable, nm_coarse_head_matches_worktree), renames the exemption comment's "the one exemption" phrasing now that a second complementary exemption exists, and points the stale FM_CREW_STATE_RUNS_LIMIT comment at fm_nm_runs_status_for_worktree (judge follow-up #1). The parent coarse-guard test's fixture is the ledger-anchored continuation shape, so its expectation flips to the fixed behavior (working via run-step, never the older failed row); a new mismatched-anchor coarse negative control preserves that guard's original no-anchor protection (pane answers, never the older row). * no-mistakes(document): Clarify pipeline attribution documentation * fix(bin): pre-register claude workspace trust at spawn time (kunchenguid#3663) * fix(bin): pre-register claude workspace trust for task worktrees A claude crewmate launched into a fresh task worktree met Claude Code's interactive workspace-trust dialog before it ever read its brief, and firstmate could not answer it: the key plane carries only Enter, Escape, and C-c with no arrow navigation, and the dialog's selection starts on "No, exit", so the documented Enter recipe ended the session instead of accepting it. Two workers wedged this way and were unblocked only by hand-seeding the trust store per path. --dangerously-skip-permissions does not cover that gate. `claude --help` records the dialog as skipped only in non-interactive mode, through -p or a non-TTY stdout, and a crewmate pane is interactive, so there is no launch flag to reach for. fm-spawn now pre-registers the worktree through bin/fm-claude-trust.sh in the existing claude branch, before the project settings that the same gate would otherwise block, and refuses the spawn when that write fails rather than launching a worker that would wedge. The scope test is the safety property and is structural rather than a path policy: the path must be a linked git worktree, sharing the spawning project's common dir, whose top level is exactly the resolved argument. Git is the ground truth, so the argument is never trusted on its own word, and a primary checkout, an unrelated repo, a worktree subdirectory, a plain directory, and a home directory are each refused rather than warned about or skipped. A treehouse or orca path prefix was deliberately avoided because treehouse's root is configurable, which would make a prefix both wrong and a new policy surface. One structural test covers both worktree providers. tests/fm-claude-trust.test.sh pins both halves, including a case where HOME is itself a valid linked worktree so the home guard is proven load-bearing rather than passing vacuously, plus the spawn-level proof that a claude spawn trusts its worktree and launches with the brief pointed at the same store. The adapter reference no longer tells a firstmate to press Enter on that dialog, and the shared trust reference now names every harness surface: which harnesses gate, which suppress at launch, which dodge the gate, which now pre-registers, and that a claude secondmate is excluded by design. The spawn fixture runs each spawn against a throwaway HOME so the suite cannot write the developer's real store, isolating through HOME rather than CLAUDE_CONFIG_DIR because the spawn forwards a set CLAUDE_CONFIG_DIR onto the launch command that launch-shape assertions read. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HNEN2GLnew27HFyfi4ms4v * fix(bin): create the staged trust store exclusively The staged store was written to a predictable pid-based path with a plain write, which follows a symlink. Where the Claude config directory is writable by another local account, that account could pre-create the path as a symlink and redirect the write into another file the launching user owns. The staged name now carries random bytes and is created with an exclusive "wx" open, so an existing path is refused outright instead of followed. The happy-path test also asserts no staged store survives the rename. The durability comment now states the residual window plainly: the readback proves the entry landed, not that it survives, because a vendor session that rewrites the whole store afterwards can still drop it and no lock closes that window when the writer is Claude itself. The worker then meets the dialog and stalls, which reaches firstmate as the ordinary stale wake rather than as silent success. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HNEN2GLnew27HFyfi4ms4v * no-mistakes(review): neutralise CDPATH in claude trust scope guard * no-mistakes(review): sandbox HOME in spawn tests, drop out-of-scope artifacts * no-mistakes(review): refuse unresolvable git dir, compact store, fix secondmate doc * no-mistakes(review): clear git env overrides, resolve symlinked store target * no-mistakes(review): degrade without node, fix Pi gate claim, record trust proof * no-mistakes(review): refuse without node, pin CLAUDE_CONFIG_DIR in spawn tests * no-mistakes(review): refuse relative config dir and concurrent store modification * no-mistakes(review): correct orca worktree claim, clean staged store on failure * no-mistakes(review): restore pretty-printed store, correct trust dialog docs * no-mistakes(review): arm trust gate before busy state to avoid orphans * no-mistakes(document): record claude trust pre-registration in its owner docs * no-mistakes(document): note orca limit for claude trust pre-registration * no-mistakes(ci): Fixed the Greptile P1 on bin/fm-spawn.sh by moving the Claude trust gate earlier rather than adding cleanup machinery. Diagnosis: Greptile reported that when Claude trust registration fails on tmux/Zellij/cmux/non-projected Herdr, the exit runs after the backend endpoint and /tmp/fm-<id> were created, and the abort trap cleans neither. The endpoint half is pre-existing, deliberate architecture — the two refusals immediately above the gate (the 60s `treehouse get` timeout at fm-spawn.sh:2550 and `validate_spawn_worktree` at :2487) also exit with the endpoint live and direct the operator with "inspect window $T"; spawn_abort_cleanup only reclaims orca endpoints (already covered via ORCA_ABORT_CLEANUP) and herdr projections. The temp-root half was genuinely introduced by this PR: the gate was placed beside the busy-state arm, ~30 lines after `mkdir -p "$TASK_TMP/gotmp"`, and fm-teardown can only find that root through `tasktmp=` in a meta record a refused spawn never publishes. Root-cause fix (smallest correct change, no new subsystem): - bin/fm-spawn.sh — moved the `claude*` trust gate from inside the busy-arm block up to the first point $WT is known, immediately after the `freshen_spawn_worktree_base` block and before TASK_TMP creation, the STATE setup, and the relaunch `clear_relaunch_harness_wiring` retirement. A refusal now leaves no temp root, no retired relaunch wiring, and no busy record; only the endpoint remains, in the same class as the two refusals just above it. - bin/fm-spawn.sh — the refusal message now ends with "inspect window $T", matching the existing convention so control/teardown can identify the endpoint. $T is set for every backend on the non-secondmate path. - bin/fm-spawn.sh:196 — header note corrected from "before any state is armed" to "before any per-task state exists". - tests/fm-claude-trust.test.sh — the existing refused-spawn test's own comment claimed "before any task state exists" but only asserted busy state. Renamed to test_refused_spawn_leaves_no_task_state and added an assertion that /tmp/fm-<id> is absent, with the task id suffixed by the test process pid so the assertion reads only this run's path (a stale /tmp/fm-refusedspawn from the fixed-id version was in fact present on this box). No assertions on implementation source bytes. Verification run locally: - The new assertion fails against the pre-fix bin/fm-spawn.sh ("not ok - a refused spawn stranded a temp root no teardown can find") and passes after — a real before/after regression proof. - tests/fm-claude-trust.test.sh: 20/20 ok. - tests/fm-backend.test.sh, fm-backend-orca, fm-control-relaunch, fm-spawn-dispatch-profile, fm-trace-context-spawn, fm-gotmp: all pass. - tests/fm-backlog-atomicity.test.sh: rc=0, 79 assertions ok. - bin/fm-lint.sh (repo's single lint owner, pinned ShellCheck 0.11.0 + actionlint 1.7.12): clean. - No /tmp/fm-refusedspawn* leftovers after the runs. Scope respected: no trust subsystem, no policy layer, no config surface, no endpoint-cleanup mechanism added; the change is an ordering move plus one error-message clause and the test that pins it. Adapter references and docs made no ordering claim, so none needed updating. Changes are left uncommitted in the worktree for the outer executor --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> * fix: restart every live second mate after updates (kunchenguid#3690) * feat(update): restart every live second mate after a successful update /updatefirstmate only restarted a second mate when that pass advanced its AGENTS.md or .agents/skills. An already-current home was skipped entirely, a bin/-only advance was steered instead, and a remote host that could not report its instruction diff was downgraded to a re-read. A running agent also freezes its launch-time wiring - turn-end hooks, harness flags, per-harness feature switches - and none of that is derivable from a file diff, so an unchanged tracked surface is not evidence the agent is already on the current behavior. Restart is now unconditional on a successful update of that home. Every live second mate the pass leaves on the target commit is restarted, whether it advanced or was already there. The safety contract is unchanged: open records are persisted before the agent is replaced, nothing is forced, stashed, or discarded, a home the pass had to skip is not restarted at all, and a mate whose runtime cannot prove a restart keeps the honest re-read path and is never reported as reloaded. bin/fm-ff-lib.sh gains a settled-state hook that fires for a home left at the base whether it advanced or was already there, and never for a skipped one; the instruction-gated hook the session-start convergence sweep uses is untouched. Regressions: fm-update pins the already-current mate into the restart set and the unprovable one into the nudge set, and fm-secondmate-restart drives both real commands end to end - an already-current home is named, persisted, and genuinely replaced with its checkout untouched, while the unprovable one keeps its running agent. * no-mistakes(document): Document unconditional secondmate restarts * fix(bin): close pending-reply decisions via resolve-key (kunchenguid#3696) * fix(bin): close reserved pending-reply keys via fm-send --resolve-key fm-send wrote answered: notes that the reserved-key fold ignores, so operator closes exited 0 while OPEN DECISIONS kept the decision open. Speak the owning library's close vocabulary on that path, and refuse when a reserved close cannot take effect. * no-mistakes(review): Safely quote manual decision-close recovery commands * no-mistakes(review): Reject unclosable overlong decision keys before sending * no-mistakes(review): Remove contract suffix from open decisions hint * no-mistakes(document): Document resolve-key line-cap refusal * fix(bin): prevent false missed-reply escalations (kunchenguid#3697) * fix(bin): stop false missed-reply escalations for same-basename self-home answers A healthy secondmate that wrote corr= to its own state/<id>.status never matched the parent channel, so recovery confirmed and the record escalated as pending-reply-missed. Make the report helper resolve the parent channel itself, skip parent-replies.status as wrong-home, put a readable sighting path on the missed line, and restatement-copy only that same-basename self-home file onto the parent channel. * no-mistakes(review): Resolve late replies before recovery escalation * no-mistakes(review): Tighten reply routing and regression coverage * no-mistakes(review): Preserve reply paths and require explicit home * no-mistakes(review): Encode wrong-home paths before persistence * no-mistakes(document): Document corrected secondmate reply routing * no-mistakes(lint): Fix pending-reply ShellCheck warnings * feat: add verified Gemini crewmate runtime (kunchenguid#3695) * feat(harness): verify gemini as a crewmate runtime adapter Adds Gemini CLI as a fourth dispatch target alongside claude, codex, and grok, scoped to crewmate and scout work only. Every axis was proven against gemini-cli 0.58.0 rather than inferred; docs/verification/runtime-backends.md carries the dated evidence and names what stayed unverified. Busy state is semantic, not rendered: BeforeAgent opens a turn and AfterAgent and SessionEnd close it. AfterAgent also fires on a manual interrupt, so a cancelled turn closes its own record. Three findings shaped the wiring rather than a config line: - --skip-trust and GEMINI_CLI_TRUST_WORKSPACE=true are presented by the CLI as equivalents and are not. A controlled A/B showed --skip-trust leaves project configuration unloaded, so workspace skills never load. - The worktree's .gemini/settings.json is the PROJECT's committed settings file, unlike claude's settings.local.json. Firstmate's hooks therefore go to a firstmate-owned state/<id>.gemini-settings.json reached through GEMINI_CLI_SYSTEM_SETTINGS_PATH, which also works untrusted and merges with a project's own hooks instead of replacing them. - The shipped CLI is a node bundle whose live process reports comm as MainThread, so ancestry cannot see it. GEMINI_CLI=1 is load-bearing and is tested before an inherited CLAUDECODE, and pane liveness identifies gemini from the script argument through the new bin/fm-gemini-lib.sh. Gemini is refused for secondmates: it has no primary supervision protocol. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * test: clear gemini's marker in launch and detection expectations Every non-gemini launch now clears GEMINI_CLI the way it already clears cursor's markers, so the two tests that pin the exact launch prefix are updated to match. The harness-detection tests that scrub foreign markers before probing ancestry scrub GEMINI_CLI too, so running the suite from inside a gemini session cannot produce a false verdict. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * docs: classify the gemini harness reference The documentation inventory is the single classification owner for maintained prose surfaces, and every surface must appear in it exactly once. The new harness reference is agent-runtime, matching its siblings. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * no-mistakes(review): Narrow Gemini ancestry detection * no-mistakes(review): Restrict Gemini hooks to canonical launches * no-mistakes(document): Document Gemini adapter support boundaries * no-mistakes(ci): Fixed Gemini process identity when interpreter or script paths contain whitespace. Tmux liveness now uses NUL-delimited /proc argv on Linux, with the existing flattened ps fallback elsewhere. Added a real-process regression test. Verified with the Gemini harness test suite, full fm-lint, ShellCheck, and git diff --check. The CI and Require no-mistakes runs were action_required/attestation outcomes rather than code failures --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(teardown): conclude parked runs advanced past task copy (kunchenguid#3704) * conclude parked runs the pipeline advanced past the task copy A no-mistakes fix round commits in the daemon's own gate-repo clone, so a run parked at a gate can carry a head whose object the task copy never received. Teardown's strict object-local identity rule then declined to conclude the run, and cleanup left it parked forever holding a fleet slot (observed 2026-09-03; the same masking condition PR 3681 fixed on the read path, now closing the teardown half its scope boundary deferred). task_status_is_own_parked_run now falls back - only when the reported head resolves to no local object - to the one shared runs-ledger attribution rule fm_nm_runs_status_for_worktree (bin/fm-nm-run-lib.sh), whose anchored continuation proof binds the branch's newest active row to this worktree's exact submitted head. Foreign branches, stale history, terminal rows, ancestor-only anchors, diverged newer rows, and ambiguous multi-row shapes all still refuse, and runs that are actively running, fixing, or in CI remain untouched: only the parked-at-a-gate determination ever reaches the abort. No sqlite access, no fetches into another task copy, no custody changes, no duplicated matching logic. * tighten the parked-run ledger fallback and pin both judge corrections The teardown ledger fallback now authorizes concluding this task's parked run only when the shared runs-ledger rule's proved answer is the explicitly active word (running): a terminal newest row - even anchored at exactly the worktree's head - is finished history and never an abort authorization. The read path may classify the same owner's answer; teardown's abort must never fire for a run that already ended. Two bounded pre-validation corrections from the implementation review: - a fetched-object counterfactual pins the strict-rule path: a pipeline fix head fetched into the task copy aborts through object-local identity alone, with an empty ledger and a proof the runs query never fired; - a negative fixture pins the tightened boundary: an unresolvable reported head with a terminal newest same-branch row anchored at the worktree head engages the ledger fallback and still refuses, so the refusal is the terminal-word boundary and not an earlier guard. * no-mistakes(review): Bind teardown ledger fallback to validated run heads * no-mistakes(review): Restore validated advanced-head ledger continuation * no-mistakes(review): Reject invalid ledger dates and terminal statuses * no-mistakes(document): Document teardown ledger scan limit * feat(bin): show requested vs effective model in Herdr agent view Track spawn-config requested_model separately from runtime-verified effective_model, probe Claude/Pi transcripts for exact API ids, push compact display metadata to Herdr, and preserve verified models across relaunch/compaction hooks without inferring aliases as truth. * fix(bin): keep re-probing effective model after first exact reading fm-model-sync.sh only probed for the runtime-verified effective model while it was still pending/UNKNOWN, so a session that later switched models (manual switch, provider fallback) kept displaying the first verified model forever and never appended a fallback-history entry. Probe unconditionally instead; fm_model_record_effective already no-ops when the probed value is unchanged, so this stays cheap. Addresses the Greptile P1 finding on PR kunchenguid#3705's fm-model-sync.sh. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): distinguish Cursor Grok, direct xAI Grok, and Anthropic Claude in the display Kapitänskorrektur: harness alone conflated Cursor-hosted Grok models (cursor-grok-4.6-*) and direct xAI Grok models (xai/grok-4.6) under one generic label, and displayed Anthropic Claude without naming the provider. Add fm_model_source_label, pattern-matched on the verified exact model id, so the compact display always reads Cursor · Grok, xAI · Grok, or Anthropic · Claude with the exact model id appended. Falls back to the existing harness label for every other model. No routing change: this only affects display strings. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): wire model-sync into the Pi extension's turn lifecycle fm-model-sync.sh was only invoked from Claude's SessionStart/ UserPromptSubmit/Stop hooks; the Pi harness's own extension (state/<id>.pi-ext.ts) never called it, so a Pi-hosted session (e.g. a pi/xai-grok crewmate) never refreshed its effective model after the first probe and Herdr kept showing the stale value with no fallback-history entry. Call fm-model-sync.sh from the same agent_start/turn_end boundaries Pi already uses for busy-state and the turn-end notification touch. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): serialize fm-model-sync.sh's meta read-probe-write Overlapping lifecycle events (Pi's agent_start/turn_end, Claude's SessionStart/UserPromptSubmit/Stop) can invoke fm-model-sync.sh concurrently for the same task. The unlocked read-probe-write let interleaved runs revert a newer effective model, mismatch its source, or duplicate a model-history entry. Serialize the critical section through the same per-task meta lock fm-spawn.sh already uses (fm_meta_lock_path + fm_lock_acquire_wait/fm_lock_release), released before every exit path. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> Co-authored-by: Arthur Haro <38157909+haroarthur@users.noreply.github.com> Co-authored-by: Nicolas Payette <nicolas.payette@specira.ai> Co-authored-by: Jon Roosevelt <rooseveltadvisors@gmail.com> Co-authored-by: att430 <41454889+att430@users.noreply.github.com> Co-authored-by: Valentino-Sole <171032438+Valentino-Sole@users.noreply.github.com>
Owner
Author
|
Closing on the captain's decision: this investigation proposed no code change and is superseded by the fixes that landed since (pi-cli-mate usage accounting, the live AGENTS.md refresh in #7, the port in #10). The branch stays as the archive. Generated by Claude Code |
Valentino-Sole
added a commit
that referenced
this pull request
Sep 8, 2026
* fix: start a fresh supervision branch for every main session (kunchenguid#3600) * fix(pi): start a new supervision branch conversation per main session The supervision branch reopened one recorded conversation forever, so every main session start reloaded the current generated prompt and then weeks of accumulated thread, where a superseded rule could still outweigh today's. The branch conversation is now scoped to one main session: the session generation owns the recorded conversation, so a cold start, /new, /resume, /fork, or a reload always builds a new one, while a rebuild inside one session (a model or effort change) still continues that session's own conversation. The dialog mirror re-anchors with it. Its durable cursor records what the previous branch conversation received, so a /resume or reload - which keeps main's own session file - would otherwise leave the new branch blind to dialog main itself still has. The reset is bounded by the current main session, and the cursor keeps advancing incrementally within it. The durable outcome store and its processed marker are untouched, so unacknowledged captain-facing outcomes still re-present on the new main session. * no-mistakes(document): Document fresh Pi supervision conversations * no-mistakes(ci): Fixed the flaky concurrent inbox failure. Lock acquisition now retries when a competing lock disappears between a failed claim and inspection. Added a behavioral regression covering that race. Verified the full inbox test four times, project lint, and git diff checks * feat: restart second mates after instruction updates (kunchenguid#3614) * feat(update): restart second mates whose instructions changed /updatefirstmate pulled new bytes onto disk and then asked each advanced second mate to re-read them. A running agent holds AGENTS.md and every loaded skill frozen from launch and no verified harness offers a reload, so that steer could not reach a loaded skill at all and left the mate holding two contradictory copies of its own job description. An eligible mate is now restarted instead, in the same home and endpoint, through the existing transactional relaunch. The restart is gated on the mate first writing down the open work it holds only in conversation - the open-record half of /stow, never its memory sweeps - so an unregistered captain call is flushed before the conversation is spent. Anything that leaves the reload unprovable falls back to the old re-read message and is reported as exactly that, never as a clean reload. Remote mates take the same path: fm-remote-secondmate-control.sh gains a relaunch verb whose host-local leg runs that same control plane, since the mate is an ordinary local secondmate from its host's point of view. The primary resolves the profile and passes it explicitly, because config/secondmate-harness is not inherited and the file on that host belongs to a different home. fm-update.sh now splits its advanced live mates into a restart set and a nudge residual, and both sets require a changed instruction surface, which also closes the over-nudge against the session-start sweep. Restart is stricter still: a bin/-only advance reloads itself on the next call, so it never costs a conversation. Colocated tests cover the gating, the persist-then-restart order, the task-subset persist request, each unsafe fallback, the remote hop, and the remote sync's new instruction-surface report. * no-mistakes(review): Fix restart correlation, concurrent waits, and lifecycle reporting * no-mistakes(review): Parallelize relaunches and classify replacement incarnations * no-mistakes(review): Gate restart actions on live agent state * no-mistakes(review): Handle failed restart workers without hanging * no-mistakes(review): Nudge legacy remotes and preserve persist recovery * no-mistakes(review): Document one-time secondmate restart rollout * no-mistakes(review): Honor arrived replies and refresh remote profiles * no-mistakes(review): Revert remote parent profile reconciliation * no-mistakes(review): Reset remote profile defaults and honor published results * no-mistakes(review): Preserve fallback nudges for unverifiable secondmates * no-mistakes(document): Document second-mate restart update flow * no-mistakes(lint): Fix ShellCheck warnings in restart scripts * perf: accelerate local validation with bounded concurrency (kunchenguid#3644) * perf(tests): route gate verification through the bounded concurrent runner Local validation was the pipeline's dominant cost: across 67 recorded no-mistakes agent sessions on this repo, 99.3% of command execution was `bash tests/*.test.sh`, run strictly one script at a time, and 2% of those calls were killed by an agent-guessed timeout and paid for twice. Three changes, each measured: - `.no-mistakes.yaml` pins `commands.test` to `bin/fm-test-run.sh --changed --exclude-family real-herdr-gated`. The runner already owns changed-file selection, bounded concurrency, the refusal of unproven scripts, and a generous automatic per-script bound, so the gate's baseline is neither a serial chain nor a guessed timeout. It stays intent-targeted - the Test step still runs its evidence agent on top - and excludes the live-Herdr family the required Herdr lane owns. - `bin/fm-test-run.sh` gives a plain list of script paths the same bounded automatic scheduler and automatic bound that `--changed` gets. Naming several subjects is how a verification round asks for exactly those scripts. The curated selections are untouched: `--lane` still composes CI shards whose serial lane must stay serial, `--family` is what the required Herdr lane runs, and `--all` stays a deliberate complete regression. - `pr-forge` is admitted to the concurrent-safe family registry on two consecutive clean proofs. `docs/fm-test-isolation-proof.md` records those, and records `secondmate` and `session-bootstrap` as refused with the exact script and reason each failed on, so the refusals are actionable rather than silent. Measured on this host, 0 failures on both sides: verification round, 4 scripts 448s chained -> 231s through the runner (-48%) pr-forge family 409.2s at 1 worker -> 237.9s at 4 (1.72x) watcher-wake-lock family 1311.1s at 1 worker -> 539.3s at 4 (2.43x) A fourth lever was implemented and then removed because the measurement refused it: raising the bounded-wait sample interval from 0.1s to 0.5s made `fm-watch-triage.test.sh` slower, 435s and 440s against 390s and 393s unchanged, back to back. Those sleeps are not overhead added to the clock - they are how a test waits for a subject moving on fm-watch.sh's own one-second cadence - so sampling less often only delays detection. It also broke `fm-watcher-lock.test.sh`, which catches a transient rather than waiting for a settled condition. CONTRIBUTING.md records that result so the experiment is not repeated. * no-mistakes(review): Separate concurrent runs by isolation proof family * no-mistakes(review): Limit automatic timeouts to changed-file validation * no-mistakes(document): Clarify validation concurrency documentation * fix: copy PR URLs from durable records (kunchenguid#3648) * fix: copy PR URLs from records or abstain, never assemble them Supervision reported a plausible but dead PR link three times because its prompt demanded a full https:// URL at a moment when only a PR number was observable, so the model assembled an owner/repository from memory, and the PR check then accepted that URL and wrote it into the task record, after which the model kept defending its own tool-endorsed guess over the worker's real link. Three changes close that chain without any live forge lookup, so private forges are treated exactly like public ones: - bin/fm-branch-prompt.sh no longer mandates a URL. Its new "PR identity: copy or abstain" section requires a URL to be copied verbatim from a durable record (the done: PR <url> status line, pr= metadata, or the backlog note), forbids assembling owner, repository, host, or number from memory, and has the branch report only the identifier it actually holds when no record names the URL yet, leaving the PR check unarmed until the worker's ready line arrives. AGENTS.md section 7 and 9 carry the same copy-or-abstain rule for main in place of the bare full-URL mandate. - Worker briefs (bin/fm-brief.sh, ship and scout rules) require the full https:// URL wherever a PR is mentioned - status line, terminal, or summary - never a bare "PR 108", so the link is in view as early as the number is. - bin/fm-pr-check.sh refuses, offline and before any side effect, a URL that the task's own done lines contradict, printing both spellings; a log naming no URL still records the argument as before. fm_pr_status_ready_urls in bin/fm-pr-lib.sh owns reading those lines. The refusal also reaches bin/fm-pr-merge.sh, so nothing merges under a contradicted URL. Tests cover the offline refusal with zero side effects, the recorded spelling being accepted, markdown-wrapped and punctuated URLs, working lines not counting, the merge wrapper propagation, a self-hosted merge request with no forge call, the prompt carrying the rule, and the brief carrying the worker rule. * no-mistakes(review): Remove stale PR URL enforcement * no-mistakes(ci): Removed backlog notes as an accepted PR identity source. PR URLs may now be copied only from the task’s `done: PR <url>` status or canonical `pr=` metadata; otherwise supervision reports only the known identifier and leaves PR checking unarmed. Updated related guidance/docs and verified with branch-supervision tests, brief tests, ShellCheck, and `git diff --check` * fix(bin): disable Claude feedback drafts for fleet launches (kunchenguid#3661) * fix(bin): disable Claude's feedback-draft flow for fleet-launched agents Scope --settings '{"feedbackDrafts":"off"}' to every Firstmate-launched Claude crewmate and secondmate, so /bug and /feedback never queue or submit a bug report on the captain's behalf. feedbackDrafts is the documented settings key (Claude Code changelog 2.1.247); the per-launch CLI flag never touches the captain's global settings.json. Claude-Session: https://claude.ai/code/session_01XYAXXzr4oZx9NjZb1veeE3 * no-mistakes(review): Prevent managed settings from re-enabling Claude feedback drafts * no-mistakes(document): Fix Claude feedback documentation formatting * fix(bin): layer both feedback-draft controls for defense in depth The prior --settings-only fix can be overridden by a managed Claude settings policy (feedbackDrafts precedence). Keep CLAUDE_CODE_SEND_FEEDBACK=0 alongside --settings '{"feedbackDrafts":"off"}': either control alone disables the SendFeedback tool, so a managed override of one still leaves the other in force. Claude-Session: https://claude.ai/code/session_01XYAXXzr4oZx9NjZb1veeE3 * no-mistakes(document): Document Claude feedback-draft suppression ownership * feat(tests): run three more validation families concurrently (kunchenguid#3662) * perf(tests): admit three more families to concurrent validation The three families that `docs/fm-test-isolation-proof.md` recorded as refused were not refused for concurrency. Each blocker was a test that decided a property by wall clock, or a script filed where it cannot run. Fixing those three things admits all three families and recovers 28.6 minutes of local validation with no assertion removed or weakened. - `tests/fm-backlog-handoff.test.sh` injected its pre-move crash by killing the handoff, sleeping a fixed second, then delegating the move to the real binary. Nothing ever killed the fake, so on a host slow enough for the case's next assertions to take longer than a second, the orphan woke and completed the very move the case requires left undone, and recovery then failed with `Task "pre-move-crash" not found in this backlog`. Watching the two backlogs during the injected crash showed exactly that, the item moving one second after the crash. All four crash injections in the file now go through a new `fm_fake_crash_injector` shim that signals the target and returns only once it is observably gone, and the pre-move fake never delegates the move at all. - `tests/fm-session-start.test.sh` proved the startup digest does not block on a slow current-state read by timing the whole digest against a fixed eight-second sleep, which a loaded host exceeds without the property being violated. It now holds that read open until the case releases it and asserts, the moment the digest returns, that the read has not finished. A digest that waited would wait indefinitely rather than for an interval a slow host can out-run, so the assertion is stronger than the bound it replaces. Its scan budget moves to the maximum, because the old value left two seconds of margin over the fixed sleep and measured the host rather than the deadline that `tests/fm-inactive-reconcile.test.sh` owns. - `fm-backend-herdr-focus-flash-e2e` was filed in the family map's catch-all, which put it in the portable serial lane, where Linux CI gate-skips it: that real-Herdr regression was running nowhere. It moves to `real-herdr-gated` and the required Herdr lane. `fm-claude-stop-autoarm-live-e2e` gate-skips on its opt-in variable and moves to `live-harness-optin`. The 28 remaining ungrouped scripts become an enumerated `standalone` family instead of admitting `unclassified` itself. `unclassified` is the family map's `*)` arm, so admitting it would silently grant concurrency to every test added afterwards, which is exactly the population with no proof. A new test still lands in `unclassified` and stays serial, and `tests/fm-test-run.test.sh` covers that split behaviorally. Each family passes two consecutive four-worker proofs with zero failures. On the production runner, `secondmate` goes 1233.1s to 453.4s, `session-bootstrap` 756.4s to 286.4s, and `standalone` 724.6s to 261.1s: 2.71x overall and 1713.2s recovered. The whole suite runs 177 scripts in 52.6 minutes of wall clock against 121 minutes of summed script time. * no-mistakes(document): Refresh concurrent validation and shard documentation * no-mistakes(ci): Fixed the real-Herdr focus-flash E2E race exposed by reclassification. Part C now starts its persistent child atomically via `pane run` and verifies stable child identity through Herdr’s public `process-info` interface, avoiding the racy send-text/send-keys sequence and platform-specific `ps` matching. Verified with bash syntax checking, ShellCheck, git diff checks, and the complete E2E test on Herdr 0.8.2 * feat: structure no-mistakes ask-user escalations (kunchenguid#3670) * feat(brief): structure no-mistakes ask-user escalation as event + snapshot file Crewmates escalating a no-mistakes ask-user gate now report one status event naming every finding id plus a snapshot file holding the gate's axi finding records verbatim (id, severity, file, line, description, authority), using the same shape even for a single finding. The status line never paraphrases. The format is defined once in fm-dod-lib.sh and rendered into both the scout and ship rule 6 in fm-brief.sh, so a promoted scout - whose rule 6 fm-promote.sh preserves unchanged - gets the identical contract as a freshly-spawned no-mistakes ship worker. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PpiWaDerbYavTLPPtEjQei * no-mistakes(review): Preserve ask-user escalation output contract * no-mistakes(review): Align escalation format test expectation * no-mistakes(review): Scope ask-user escalation instructions correctly * no-mistakes(review): Remove ask-user from generic decision rules --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> * fix(bin): require self-sufficient no-mistakes intent (kunchenguid#3671) * fix(bin): require a self-sufficient no-mistakes intent A no-mistakes worker's --intent is only as useful as the string it passes. PR kunchenguid#3604 shipped with an intent that was only "do 1, 2, 3, 7 from the report": the real contract lived in a private scout report and never reached --intent, so nobody holding that string plus the codebase could have derived the specification. This is pure instruction at the contract's one owner; no spawn-side or promotion-side check is added. - bin/fm-dod-lib.sh: the generated no-mistakes Definition of done now states that the --intent string must be self-sufficient (the string plus the codebase reconstructs roughly the same specification) and tells the worker to write the substance of any report, decision, or PR the captain's intent refers to into --intent rather than the pointer, while Firstmate build instructions and the worker's own decisions still stay out. The spawn-time overlay points back at that rule so its "supersedes" wording cannot cancel it, and the header's owner statement carries the rule. - AGENTS.md section 11 and bin/fm-brief.sh's header ask Firstmate to include the substance of referenced material when filling ## Captain's intent, and section 11 points at the owner of the rule. - tests/fm-brief.test.sh and tests/fm-task-delivery.test.sh assert the rendered brief and launch contract carry the rule. Claude-Session: https://claude.ai/code/session_01YMhEe42q7BAAoN6RxNuzim * no-mistakes(document): Replace incident-specific intent test commentary * fix: accelerate local Bearings snapshot composition (kunchenguid#3499) * Speed local fleet snapshot composition * no-mistakes(review): Stabilize task inventory during concurrent snapshot composition * no-mistakes(document): Document local snapshot observation concurrency * no-mistakes(ci): Fixed CI failures by making empty task manifests compatible with stock macOS Bash 3.2, snapshotting task metadata before concurrent observations to prevent generation drift, strengthening the behavioral race regression, and updating the stock-Bash Bearings test count to 45. Verified fleet snapshot tests (15), Bearings tests (45), workflow lint tests, project lint, Bash 3.2 parsing, and diff checks * no-mistakes(ci): Fixed the Linux CI failure caused by passing large backlog/task JSON through jq command-line arguments, which exceeded the per-argument size limit. Both inventory projections now stream large JSON inputs through stdin. Verified with fm-bearings-snapshot.test.sh (45 tests), fm-fleet-snapshot-view.test.sh (15 tests), Bash syntax, and git diff checks * no-mistakes(ci): Fixed concurrent task teardown during metadata capture: vanished metadata is now omitted while genuine copy failures remain fatal. Added a deterministic public Bearings regression test and updated CI’s expected test count. Verified with the full Bearings suite, workflow-lint suite, Bash syntax checks, and git diff checks * no-mistakes(ci): Fixed PR-caused CI and review issues: streamed large fleet JSON through jq stdin to avoid Linux argument limits, kept crew-state reads bound to captured metadata generations, and strengthened the behavioral race test. Bearings (46 tests), fleet snapshot (15 tests), crew-state, backend, lint, Bash syntax, and diff checks pass locally. Serial shard 5’s unrelated task-inbox segmentation fault appears infrastructural/flaky * no-mistakes(ci): Fixed endpoint-state generation crossing by validating captured spawn_gen before and after local endpoint probes, falling back to exact metadata identity for legacy tasks. Stale probe results now become unknown instead of false unhealthy state. Added a behavioral relaunch-race regression test. Verified the full Bearings snapshot suite, shellcheck, bash syntax, and git diff checks * fix(snapshot): keep live observations generation-coherent * no-mistakes(review): Keep secondmate observations generation-bound without copying reports * no-mistakes(document): Document generation-coherent snapshot observations * test(bearings): measure local read overlap instead of wall-clock budget The large-local-snapshot regression asserted that a whole snapshot composed in under five seconds. That bound measures how loaded the host is, not whether the per-task reads actually overlap, so it failed intermittently on a contended machine: one run in six on a box at load 16-20, landing exactly on the five second boundary. Time a serialized run and a concurrent run of the same workload instead and require the concurrent one to save at least two seconds. Both runs pay the same composition overhead, so the difference isolates the overlap this change delivers. Five one-second reads serialize into five seconds and overlap into about one, and re-serializing the reads collapses the saving to roughly zero, so the assertion still fails loudly if the concurrency regresses. Also bump the pinned Bearings test count to 48, since rebasing onto the current default branch picked up its captain-hold test. * no-mistakes(review): Restore JSON-derived decision flags * no-mistakes(review): Unify status-derived snapshot observations * no-mistakes(ci): Updated the stock macOS Bash CI check’s Bearings test count from 48 to 49. Verified the full Bearings suite passes and emits exactly 49 TAP successes; git diff checks pass * fix: prevent stale supervision wake loops (kunchenguid#3672) * fix(bin): stop the supervision branch's stale-ack and ghost-report loops Clean-slate implementation of the four authorized recommendations from the supervision-ghost-retrigger analysis (items 1, 2, 3, and 7), in their minimal form, superseding PR kunchenguid#3604: - fm_branch_report refuses a task the wake being handled never named. The extension fixes the reportable task set from the eligible rows before each prompt (signal and stale rows resolve to their tasks, a heartbeat allows any task with a live record, fleet is always allowed), so a report typed from memory about a task whose records teardown already removed is never stored or delivered. - An acknowledgement that consumes nothing says "nothing was acknowledged through N" and prints the exact --ack-through / --recovery-generation command for the current presented wake, instead of "re-run the drain", which re-fed the same stale acknowledgement in a loop. - bin/fm-guard.sh no longer tells the branch actor to drain queued wakes while it is handling them; it names the granted rows instead. - Teardown removes state/.<task>.branch-outcome-index for ordinary tasks and descendants; the index rebuild and the append-side index write both skip a task with neither a live record nor a status log, so the branch's report of a teardown it just performed is stored without recreating the index. No new locking, no spawn-generation binding, and no retired-task refusal: the branch can still report the outcome of a task it just tore down, and the teardown test now proves that path end to end. * fix(bin): narrow the branch report scope and guard silence to the minimal form Apply the four review decisions on the clean-slate branch: - A signal or stale prompt may report only the tasks its own rows resolve to; fleet is refused there too. A heartbeat review is not scoped by task at all, so the extension no longer tracks live task records and refuses nothing by task id during a fleet review. - The outcome-index rebuild no longer skips retired tasks; the append-side skip alone keeps a torn-down task's index from being recreated. - bin/fm-guard.sh keeps the queued-wakes warning silent for the branch actor instead of printing a replacement note. * no-mistakes(document): Align supervision docs with scoped wake handling * fix(bin): avoid fleet snapshot argument limits (kunchenguid#3677) * Fix fleet snapshot large JSON transport * no-mistakes(review): Captain: file-back fleet snapshot transport safely * no-mistakes(review): Captain: file-back parent summary aggregation * no-mistakes(ci): Rebased the PR's three commits onto f4d7875 and resolved the fleet snapshot conflict while preserving the base's task-observation lifecycle. Fixed Greptile's valid finding by recursively removing the private mktemp transport directory, so future transport files cannot cause cleanup to fail. Verified with tests/fm-home-summary-refresh.test.sh, bin/fm-lint.sh, git diff --check, and ancestry checks. All passed; the fix remains as an uncommitted worktree change for the outer executor * fix(bin): attribute active runs with unfetched pipeline heads (kunchenguid#3681) * fix(bin): recognize active pipeline fix rounds with unfetched run heads A no-mistakes fix round advances the run head beyond the submitted head, and the pipeline commits in its own checkout, so the task copy never receives the new commit object. fm-crew-state's strict head rule rejected the active row, the coarse runs-list scan skipped it and matched the older failed row at the submitted head, and an active validation read as failed (observed on model-routing-benchmark-hardening: active head ac61c64 vs task copy at fb47636d). fm_nm_runs_status_for_worktree in bin/fm-nm-run-lib.sh now owns runs-ledger attribution: the branch's newest row alone decides, and a newest row whose head cannot resolve locally is recognized only as a provable pipeline-owned continuation - active (running) and anchored by the immediately older row for the same branch having ended at exactly this worktree's HEAD. The reader keeps the axi TOON as full detail for that proven same-branch run. Unanchored, ancestor-anchored, and terminal unresolvable rows stay unattributed, so branch-name coincidence and other tasks' runs never match, and fm_nm_head_matches_worktree keeps its exact prior semantics for teardown (verified by the full teardown suite). Tests: reproduction regression for the unfetched active fix head (reads working via full run-step detail), coarse-path continuation when axi answers another branch, and negative controls for the unanchored active row and the unresolvable terminal row with the historical fallback preserved. Ported onto upstream/main f4d7875, where kunchenguid#3194 independently added the branch_sync custody exemption on the full axi-status path: both mechanisms now coexist, each owning one surface (TOON custody on the full path, the runs ledger on the coarse path). The port deletes the superseded coarse scan-and-skip (nm_runs_status_for_branch) and its now caller-less helpers (fm_nm_head_resolvable, nm_coarse_head_matches_worktree), renames the exemption comment's "the one exemption" phrasing now that a second complementary exemption exists, and points the stale FM_CREW_STATE_RUNS_LIMIT comment at fm_nm_runs_status_for_worktree (judge follow-up #1). The parent coarse-guard test's fixture is the ledger-anchored continuation shape, so its expectation flips to the fixed behavior (working via run-step, never the older failed row); a new mismatched-anchor coarse negative control preserves that guard's original no-anchor protection (pane answers, never the older row). * no-mistakes(document): Clarify pipeline attribution documentation * fix(bin): pre-register claude workspace trust at spawn time (kunchenguid#3663) * fix(bin): pre-register claude workspace trust for task worktrees A claude crewmate launched into a fresh task worktree met Claude Code's interactive workspace-trust dialog before it ever read its brief, and firstmate could not answer it: the key plane carries only Enter, Escape, and C-c with no arrow navigation, and the dialog's selection starts on "No, exit", so the documented Enter recipe ended the session instead of accepting it. Two workers wedged this way and were unblocked only by hand-seeding the trust store per path. --dangerously-skip-permissions does not cover that gate. `claude --help` records the dialog as skipped only in non-interactive mode, through -p or a non-TTY stdout, and a crewmate pane is interactive, so there is no launch flag to reach for. fm-spawn now pre-registers the worktree through bin/fm-claude-trust.sh in the existing claude branch, before the project settings that the same gate would otherwise block, and refuses the spawn when that write fails rather than launching a worker that would wedge. The scope test is the safety property and is structural rather than a path policy: the path must be a linked git worktree, sharing the spawning project's common dir, whose top level is exactly the resolved argument. Git is the ground truth, so the argument is never trusted on its own word, and a primary checkout, an unrelated repo, a worktree subdirectory, a plain directory, and a home directory are each refused rather than warned about or skipped. A treehouse or orca path prefix was deliberately avoided because treehouse's root is configurable, which would make a prefix both wrong and a new policy surface. One structural test covers both worktree providers. tests/fm-claude-trust.test.sh pins both halves, including a case where HOME is itself a valid linked worktree so the home guard is proven load-bearing rather than passing vacuously, plus the spawn-level proof that a claude spawn trusts its worktree and launches with the brief pointed at the same store. The adapter reference no longer tells a firstmate to press Enter on that dialog, and the shared trust reference now names every harness surface: which harnesses gate, which suppress at launch, which dodge the gate, which now pre-registers, and that a claude secondmate is excluded by design. The spawn fixture runs each spawn against a throwaway HOME so the suite cannot write the developer's real store, isolating through HOME rather than CLAUDE_CONFIG_DIR because the spawn forwards a set CLAUDE_CONFIG_DIR onto the launch command that launch-shape assertions read. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HNEN2GLnew27HFyfi4ms4v * fix(bin): create the staged trust store exclusively The staged store was written to a predictable pid-based path with a plain write, which follows a symlink. Where the Claude config directory is writable by another local account, that account could pre-create the path as a symlink and redirect the write into another file the launching user owns. The staged name now carries random bytes and is created with an exclusive "wx" open, so an existing path is refused outright instead of followed. The happy-path test also asserts no staged store survives the rename. The durability comment now states the residual window plainly: the readback proves the entry landed, not that it survives, because a vendor session that rewrites the whole store afterwards can still drop it and no lock closes that window when the writer is Claude itself. The worker then meets the dialog and stalls, which reaches firstmate as the ordinary stale wake rather than as silent success. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HNEN2GLnew27HFyfi4ms4v * no-mistakes(review): neutralise CDPATH in claude trust scope guard * no-mistakes(review): sandbox HOME in spawn tests, drop out-of-scope artifacts * no-mistakes(review): refuse unresolvable git dir, compact store, fix secondmate doc * no-mistakes(review): clear git env overrides, resolve symlinked store target * no-mistakes(review): degrade without node, fix Pi gate claim, record trust proof * no-mistakes(review): refuse without node, pin CLAUDE_CONFIG_DIR in spawn tests * no-mistakes(review): refuse relative config dir and concurrent store modification * no-mistakes(review): correct orca worktree claim, clean staged store on failure * no-mistakes(review): restore pretty-printed store, correct trust dialog docs * no-mistakes(review): arm trust gate before busy state to avoid orphans * no-mistakes(document): record claude trust pre-registration in its owner docs * no-mistakes(document): note orca limit for claude trust pre-registration * no-mistakes(ci): Fixed the Greptile P1 on bin/fm-spawn.sh by moving the Claude trust gate earlier rather than adding cleanup machinery. Diagnosis: Greptile reported that when Claude trust registration fails on tmux/Zellij/cmux/non-projected Herdr, the exit runs after the backend endpoint and /tmp/fm-<id> were created, and the abort trap cleans neither. The endpoint half is pre-existing, deliberate architecture — the two refusals immediately above the gate (the 60s `treehouse get` timeout at fm-spawn.sh:2550 and `validate_spawn_worktree` at :2487) also exit with the endpoint live and direct the operator with "inspect window $T"; spawn_abort_cleanup only reclaims orca endpoints (already covered via ORCA_ABORT_CLEANUP) and herdr projections. The temp-root half was genuinely introduced by this PR: the gate was placed beside the busy-state arm, ~30 lines after `mkdir -p "$TASK_TMP/gotmp"`, and fm-teardown can only find that root through `tasktmp=` in a meta record a refused spawn never publishes. Root-cause fix (smallest correct change, no new subsystem): - bin/fm-spawn.sh — moved the `claude*` trust gate from inside the busy-arm block up to the first point $WT is known, immediately after the `freshen_spawn_worktree_base` block and before TASK_TMP creation, the STATE setup, and the relaunch `clear_relaunch_harness_wiring` retirement. A refusal now leaves no temp root, no retired relaunch wiring, and no busy record; only the endpoint remains, in the same class as the two refusals just above it. - bin/fm-spawn.sh — the refusal message now ends with "inspect window $T", matching the existing convention so control/teardown can identify the endpoint. $T is set for every backend on the non-secondmate path. - bin/fm-spawn.sh:196 — header note corrected from "before any state is armed" to "before any per-task state exists". - tests/fm-claude-trust.test.sh — the existing refused-spawn test's own comment claimed "before any task state exists" but only asserted busy state. Renamed to test_refused_spawn_leaves_no_task_state and added an assertion that /tmp/fm-<id> is absent, with the task id suffixed by the test process pid so the assertion reads only this run's path (a stale /tmp/fm-refusedspawn from the fixed-id version was in fact present on this box). No assertions on implementation source bytes. Verification run locally: - The new assertion fails against the pre-fix bin/fm-spawn.sh ("not ok - a refused spawn stranded a temp root no teardown can find") and passes after — a real before/after regression proof. - tests/fm-claude-trust.test.sh: 20/20 ok. - tests/fm-backend.test.sh, fm-backend-orca, fm-control-relaunch, fm-spawn-dispatch-profile, fm-trace-context-spawn, fm-gotmp: all pass. - tests/fm-backlog-atomicity.test.sh: rc=0, 79 assertions ok. - bin/fm-lint.sh (repo's single lint owner, pinned ShellCheck 0.11.0 + actionlint 1.7.12): clean. - No /tmp/fm-refusedspawn* leftovers after the runs. Scope respected: no trust subsystem, no policy layer, no config surface, no endpoint-cleanup mechanism added; the change is an ordering move plus one error-message clause and the test that pins it. Adapter references and docs made no ordering claim, so none needed updating. Changes are left uncommitted in the worktree for the outer executor --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> * fix: restart every live second mate after updates (kunchenguid#3690) * feat(update): restart every live second mate after a successful update /updatefirstmate only restarted a second mate when that pass advanced its AGENTS.md or .agents/skills. An already-current home was skipped entirely, a bin/-only advance was steered instead, and a remote host that could not report its instruction diff was downgraded to a re-read. A running agent also freezes its launch-time wiring - turn-end hooks, harness flags, per-harness feature switches - and none of that is derivable from a file diff, so an unchanged tracked surface is not evidence the agent is already on the current behavior. Restart is now unconditional on a successful update of that home. Every live second mate the pass leaves on the target commit is restarted, whether it advanced or was already there. The safety contract is unchanged: open records are persisted before the agent is replaced, nothing is forced, stashed, or discarded, a home the pass had to skip is not restarted at all, and a mate whose runtime cannot prove a restart keeps the honest re-read path and is never reported as reloaded. bin/fm-ff-lib.sh gains a settled-state hook that fires for a home left at the base whether it advanced or was already there, and never for a skipped one; the instruction-gated hook the session-start convergence sweep uses is untouched. Regressions: fm-update pins the already-current mate into the restart set and the unprovable one into the nudge set, and fm-secondmate-restart drives both real commands end to end - an already-current home is named, persisted, and genuinely replaced with its checkout untouched, while the unprovable one keeps its running agent. * no-mistakes(document): Document unconditional secondmate restarts * fix(bin): close pending-reply decisions via resolve-key (kunchenguid#3696) * fix(bin): close reserved pending-reply keys via fm-send --resolve-key fm-send wrote answered: notes that the reserved-key fold ignores, so operator closes exited 0 while OPEN DECISIONS kept the decision open. Speak the owning library's close vocabulary on that path, and refuse when a reserved close cannot take effect. * no-mistakes(review): Safely quote manual decision-close recovery commands * no-mistakes(review): Reject unclosable overlong decision keys before sending * no-mistakes(review): Remove contract suffix from open decisions hint * no-mistakes(document): Document resolve-key line-cap refusal * fix(bin): prevent false missed-reply escalations (kunchenguid#3697) * fix(bin): stop false missed-reply escalations for same-basename self-home answers A healthy secondmate that wrote corr= to its own state/<id>.status never matched the parent channel, so recovery confirmed and the record escalated as pending-reply-missed. Make the report helper resolve the parent channel itself, skip parent-replies.status as wrong-home, put a readable sighting path on the missed line, and restatement-copy only that same-basename self-home file onto the parent channel. * no-mistakes(review): Resolve late replies before recovery escalation * no-mistakes(review): Tighten reply routing and regression coverage * no-mistakes(review): Preserve reply paths and require explicit home * no-mistakes(review): Encode wrong-home paths before persistence * no-mistakes(document): Document corrected secondmate reply routing * no-mistakes(lint): Fix pending-reply ShellCheck warnings * feat: add verified Gemini crewmate runtime (kunchenguid#3695) * feat(harness): verify gemini as a crewmate runtime adapter Adds Gemini CLI as a fourth dispatch target alongside claude, codex, and grok, scoped to crewmate and scout work only. Every axis was proven against gemini-cli 0.58.0 rather than inferred; docs/verification/runtime-backends.md carries the dated evidence and names what stayed unverified. Busy state is semantic, not rendered: BeforeAgent opens a turn and AfterAgent and SessionEnd close it. AfterAgent also fires on a manual interrupt, so a cancelled turn closes its own record. Three findings shaped the wiring rather than a config line: - --skip-trust and GEMINI_CLI_TRUST_WORKSPACE=true are presented by the CLI as equivalents and are not. A controlled A/B showed --skip-trust leaves project configuration unloaded, so workspace skills never load. - The worktree's .gemini/settings.json is the PROJECT's committed settings file, unlike claude's settings.local.json. Firstmate's hooks therefore go to a firstmate-owned state/<id>.gemini-settings.json reached through GEMINI_CLI_SYSTEM_SETTINGS_PATH, which also works untrusted and merges with a project's own hooks instead of replacing them. - The shipped CLI is a node bundle whose live process reports comm as MainThread, so ancestry cannot see it. GEMINI_CLI=1 is load-bearing and is tested before an inherited CLAUDECODE, and pane liveness identifies gemini from the script argument through the new bin/fm-gemini-lib.sh. Gemini is refused for secondmates: it has no primary supervision protocol. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * test: clear gemini's marker in launch and detection expectations Every non-gemini launch now clears GEMINI_CLI the way it already clears cursor's markers, so the two tests that pin the exact launch prefix are updated to match. The harness-detection tests that scrub foreign markers before probing ancestry scrub GEMINI_CLI too, so running the suite from inside a gemini session cannot produce a false verdict. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * docs: classify the gemini harness reference The documentation inventory is the single classification owner for maintained prose surfaces, and every surface must appear in it exactly once. The new harness reference is agent-runtime, matching its siblings. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GYdQkKfSEQrFUxtcTXZ66L * no-mistakes(review): Narrow Gemini ancestry detection * no-mistakes(review): Restrict Gemini hooks to canonical launches * no-mistakes(document): Document Gemini adapter support boundaries * no-mistakes(ci): Fixed Gemini process identity when interpreter or script paths contain whitespace. Tmux liveness now uses NUL-delimited /proc argv on Linux, with the existing flattened ps fallback elsewhere. Added a real-process regression test. Verified with the Gemini harness test suite, full fm-lint, ShellCheck, and git diff --check. The CI and Require no-mistakes runs were action_required/attestation outcomes rather than code failures --------- Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> * fix(teardown): conclude parked runs advanced past task copy (kunchenguid#3704) * conclude parked runs the pipeline advanced past the task copy A no-mistakes fix round commits in the daemon's own gate-repo clone, so a run parked at a gate can carry a head whose object the task copy never received. Teardown's strict object-local identity rule then declined to conclude the run, and cleanup left it parked forever holding a fleet slot (observed 2026-09-03; the same masking condition PR 3681 fixed on the read path, now closing the teardown half its scope boundary deferred). task_status_is_own_parked_run now falls back - only when the reported head resolves to no local object - to the one shared runs-ledger attribution rule fm_nm_runs_status_for_worktree (bin/fm-nm-run-lib.sh), whose anchored continuation proof binds the branch's newest active row to this worktree's exact submitted head. Foreign branches, stale history, terminal rows, ancestor-only anchors, diverged newer rows, and ambiguous multi-row shapes all still refuse, and runs that are actively running, fixing, or in CI remain untouched: only the parked-at-a-gate determination ever reaches the abort. No sqlite access, no fetches into another task copy, no custody changes, no duplicated matching logic. * tighten the parked-run ledger fallback and pin both judge corrections The teardown ledger fallback now authorizes concluding this task's parked run only when the shared runs-ledger rule's proved answer is the explicitly active word (running): a terminal newest row - even anchored at exactly the worktree's head - is finished history and never an abort authorization. The read path may classify the same owner's answer; teardown's abort must never fire for a run that already ended. Two bounded pre-validation corrections from the implementation review: - a fetched-object counterfactual pins the strict-rule path: a pipeline fix head fetched into the task copy aborts through object-local identity alone, with an empty ledger and a proof the runs query never fired; - a negative fixture pins the tightened boundary: an unresolvable reported head with a terminal newest same-branch row anchored at the worktree head engages the ledger fallback and still refuses, so the refusal is the terminal-word boundary and not an earlier guard. * no-mistakes(review): Bind teardown ledger fallback to validated run heads * no-mistakes(review): Restore validated advanced-head ledger continuation * no-mistakes(review): Reject invalid ledger dates and terminal statuses * no-mistakes(document): Document teardown ledger scan limit * feat(bin): show requested vs effective model in Herdr agent view Track spawn-config requested_model separately from runtime-verified effective_model, probe Claude/Pi transcripts for exact API ids, push compact display metadata to Herdr, and preserve verified models across relaunch/compaction hooks without inferring aliases as truth. * fix(bin): keep re-probing effective model after first exact reading fm-model-sync.sh only probed for the runtime-verified effective model while it was still pending/UNKNOWN, so a session that later switched models (manual switch, provider fallback) kept displaying the first verified model forever and never appended a fallback-history entry. Probe unconditionally instead; fm_model_record_effective already no-ops when the probed value is unchanged, so this stays cheap. Addresses the Greptile P1 finding on PR kunchenguid#3705's fm-model-sync.sh. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): distinguish Cursor Grok, direct xAI Grok, and Anthropic Claude in the display Kapitänskorrektur: harness alone conflated Cursor-hosted Grok models (cursor-grok-4.6-*) and direct xAI Grok models (xai/grok-4.6) under one generic label, and displayed Anthropic Claude without naming the provider. Add fm_model_source_label, pattern-matched on the verified exact model id, so the compact display always reads Cursor · Grok, xAI · Grok, or Anthropic · Claude with the exact model id appended. Falls back to the existing harness label for every other model. No routing change: this only affects display strings. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): wire model-sync into the Pi extension's turn lifecycle fm-model-sync.sh was only invoked from Claude's SessionStart/ UserPromptSubmit/Stop hooks; the Pi harness's own extension (state/<id>.pi-ext.ts) never called it, so a Pi-hosted session (e.g. a pi/xai-grok crewmate) never refreshed its effective model after the first probe and Herdr kept showing the stale value with no fallback-history entry. Call fm-model-sync.sh from the same agent_start/turn_end boundaries Pi already uses for busy-state and the turn-end notification touch. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm * fix(bin): serialize fm-model-sync.sh's meta read-probe-write Overlapping lifecycle events (Pi's agent_start/turn_end, Claude's SessionStart/UserPromptSubmit/Stop) can invoke fm-model-sync.sh concurrently for the same task. The unlocked read-probe-write let interleaved runs revert a newer effective model, mismatch its source, or duplicate a model-history entry. Serialize the critical section through the same per-task meta lock fm-spawn.sh already uses (fm_meta_lock_path + fm_lock_acquire_wait/fm_lock_release), released before every exit path. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BwfFjeYQcz9cZ3vEZohmpm --------- Co-authored-by: Kun Chen <3233006+kunchenguid@users.noreply.github.com> Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> Co-authored-by: Arthur Haro <38157909+haroarthur@users.noreply.github.com> Co-authored-by: Nicolas Payette <nicolas.payette@specira.ai> Co-authored-by: Jon Roosevelt <rooseveltadvisors@gmail.com> Co-authored-by: att430 <41454889+att430@users.noreply.github.com> Co-authored-by: Valentino-Sole <171032438+Valentino-Sole@users.noreply.github.com>
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
9 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.
Summary
getContextUsage()display clamp), and a live-discovered regression where an independentpipackage reinstall mid-investigation silently reverted both fixes.claude-code-cli/sonnet, no cross-turn cache reuse) as the actual driver of real, legitimate token growth.Delivery: push + PR only, per explicit captain instruction - no no-mistakes pipeline, no merge.
Test plan
bin/fm-doc-audience-check.shpasses (surfaces=90, local_links=302)test/agent-session-stats.test.ts,test/compaction.test.ts,test/agent-session-compaction.test.ts,test/suite/agent-session-compaction.test.ts,test/suite/regressions/8328-zero-usage-auto-compaction.test.ts,test/suite/regressions/post-compaction-error-fallback.test.ts): 69 passed, 7 intentionally skipped, 0 failedbin/fm-test-run.sh --changed --base main: 176 total, 81 passed, 71 failed, 24 gate-skipped - failures traced to session-bootstrap/secondmate environment contamination (inheritedFM_ROOT_OVERRIDE/FM_STATE_OVERRIDE) unrelated to this docs-only diff (confirmed: none of the failing tests read the changed file; clearing those env vars for the previously-11/11-failing session-bootstrap family dropped it to 8/11)🤖 Generated with Claude Code
https://claude.ai/code/session_013XPDkmktpahrVQuhPqXow5