fix(kanban): preserve initially blocked human gates

[salimhamed]

Problem

_has_sticky_block decides whether a blocked task is a deliberate human gate
(which recompute_ready must not auto-promote) by inspecting the task's most
recent blocked/unblocked event. But a task created with
--initial-status blocked — the canonical way to stand up an immediate human
gate — emits only a created event (payload status="blocked"); it has no
blocked event yet.

So such a gate was treated as non-sticky. With no unfinished parents, the empty
parent set is vacuously satisfied and recompute_ready promotes the gate to
ready; the dispatcher then claims it and spawns a worker on a card that was
meant to wait for a human.

Symptom

An initially-blocked human gate is silently promoted → dispatched → and if the
spawned worker re-blocks for input, the same-cause re-block trips
BLOCK_RECURRENCE_LIMIT → block_loop_detected → the card is routed out and
archived — instead of simply sitting blocked awaiting a human.

Fix

Treat a created event whose payload records status="blocked" as a sticky
block, so recompute_ready leaves it blocked until an explicit unblock.
Behavior for circuit-breaker/direct-DB blocks (no explicit human-block signal)
is unchanged. Adds a regression test.

[teknium1]

Thanks for the focused regression fix. The premise is present on current origin/main: create_task() records created with payload.status at hermes_cli/kanban_db.py:2670-2683, but _has_sticky_block() only examines blocked/unblocked at hermes_cli/kanban_db.py:3273-3279. recompute_ready() then considers the blocked task for promotion at hermes_cli/kanban_db.py:3317-3365.

The proposed event ordering preserves explicit unblock_task() semantics, which emits unblocked at hermes_cli/kanban_db.py:5110-5113, while retaining the existing circuit-breaker/direct-write recovery behavior. The new regression test exercises the parentless, initially blocked path directly.

Automated hermes-sweeper review.

[GottZ]

This was generated by AI during triage.

Summary

Twenty-eight PRs address or reference the two issues: most repair the missing create-time blocked lifecycle event that lets recompute_ready() release and dispatch an approval gate, while others broaden the scope into authorization, stale-state recovery, archival, promotion, or body updates. The focused event-emission diffs match the reported cause while preserving circuit-breaker recovery.

Related pull requests

• feat(kanban): add initial status for human-ops cards #27526 [closed] related — (+88/-10) — n/a: Introduced initial_status across the database, CLI, and tools but emitted only created; closed after the focused feature was salvaged into feat(kanban): add initial-status for human-ops cards (#27526) #28414.
• feat(kanban): add initial-status for human-ops cards (#27526) #28414 [merged] related — (+40/-8) — n/a: Merged reference implementation of initial_status; it remains relevant because its create path records blocked status without the lifecycle event consumed by _has_sticky_block().
• fix: keep blocked kanban tasks blocked #32581 partial — (+96/-117) — n/a: Prevents the bypass by excluding every blocked task from recomputation, but the keep_open review documents that this reverses intentional circuit-breaker recovery, so only its regression coverage is salvageable.
• fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 best fix — (+123/-0) — n/a: Recorded best fix for both issues; atomically appends blocked after created and covers repeated recomputation, explicit unblock, and done-parent behavior without changing circuit-breaker recovery.
• fix: keep explicitly blocked Kanban tasks sticky #34756 related — (+190/-10) — n/a: Adds stale sticky-row repair and a claim-boundary guard, but the contributor review says it would undo supported manual promotion and must retain current retry-limit behavior.
• fix: keep initial blocked kanban tasks inert #35832 fixes — (+354/-4) — n/a: Contains the same create-time event repair plus archive/worker teardown hardening; the contributor review blocks the bundled teardown because Windows may signal an unverified recycled PID.
• fix: keep initially blocked kanban tasks sticky #39085 fixes — (+34/-0) — n/a: Adds the same create-time blocked event and recomputation regression as fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 from an older base; despite its keep_open review, the production correction is a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): keep initial blocked tasks sticky #39832 [closed] duplicate — (+67/-12) — n/a: Infers stickiness from created(status=blocked) instead of emitting the established lifecycle event, bypassing notification and diagnostic consumers; closed in favor of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): keep initial blocked holds sticky #41497 fixes — (+94/-0) — n/a: Adds the same event as fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 plus recomputation and dispatcher dry-run tests; despite its keep_open review and contributor LGTM, its production correction duplicates fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): keep initial blocked tasks sticky #43479 [closed] fixes — (+4591/-96) — n/a: Contains the correct event insertion among extensive unrelated gateway, Discord, memory, archival, and repair changes; closed because the bundled work was not ready for public upstream due diligence.
• fix(kanban): preserve initial blocked approval gates #46274 [closed] fixes — (+66/-0) — n/a: Focused create-time event implementation with parentless and done-parent regressions; closed but still relevant as a duplicate reference for fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): keep initial_status=blocked tasks blocked across recompute_ready #46565 fixes — (+26/-0) — n/a: Appends the missing event in the creation transaction and tests recomputation; despite its keep_open review, contributor triage explicitly identifies it as a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix: keep initial blocked kanban tasks sticky #48253 [closed] duplicate — (+55/-5) — n/a: Uses created-payload inference rather than emitting blocked; closed as a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 after focused validation.
• fix(kanban): emit blocked event for initial_status="blocked" tasks (fixes #47777) #48437 [closed] fixes — (+61/-0) — n/a: Adds the same lifecycle event with repeated-recompute, done-parent, and unblock tests; closed as a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 despite a non-contributor APPROVED review.
• fix: keep initially blocked kanban cards sticky #60100 [closed] partial — (+45/-0) — n/a: Emits the event but records recurrences: 1 while persisted recurrence state remains 0; closed in favor of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 after that review finding.
• fix(kanban): keep initially blocked tasks sticky #61890 [closed] fixes — (+39/-0) — n/a: Atomically emits the same lifecycle event with done-parent coverage; closed after contributor consolidation as a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): preserve initially blocked human gates #64830 partial — (+37/-12) — n/a: Makes legacy created(status=blocked) rows sticky by changing _has_sticky_block(), stopping promotion but not producing the lifecycle event used by notifications and diagnostics.
• fix(kanban): keep initial blocked tasks sticky #66993 fixes — (+89/-0) — n/a: Adds the create-time event plus atomic rollback and manual-promotion release behavior, so its promotion semantics are competing work that should be split or independently reviewed.
• fix(kanban): enforce auditable authorization for human gates #69531 best fix — (+2451/-127) — n/a: Recorded best fix for [Bug]: Tasks created with --initial-status blocked auto-promote to ready ~1s later with no actor — human approval gate bypassed #39609 and adds authenticated, fingerprint-bound authorization through claim and spawn; the visible diff now includes provider_override, but the contributor review still requires reconciliation against current main.
• fix(kanban): keep initial blocked tasks sticky #69679 partial — (+127/-13) — n/a: Adds typed creation metadata, legacy fallback, and manual-promotion release, but the contributor review blocks recurrence initialization at 1 because the first later same-kind block would escalate instead of re-stick.
• fix(kanban): make creation-time blocks sticky #70733 fixes — (+194/-0) — n/a: Its atomic create-time event fixes the root cause, but the additional annotation path matches all blocked rows and would make recoverable circuit-breaker gave_up rows sticky, as the contributor review explains.
• fix(kanban): keep create-time blocked tasks sticky #71147 partial — (+152/-2) — n/a: Adds the event plus a narrowly conditioned idempotent legacy-row backfill; the keep_open review requires manual placement of its tests after current-main pruning.
• Keep initially blocked Kanban tasks sticky #71861 partial — (+39/-0) — n/a: Adds the same event and non-promotion regression as fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735; despite its keep_open review, contributor triage identifies it as a duplicate and the diff adds no distinct production correction.
• fix(kanban): make --initial-status blocked sticky against recompute_ready #71977 fixes — (+101/-0) — n/a: Adds the same event as fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 with dispatcher, unblock, and done-parent regressions; despite its keep_open review, contributor triage explicitly marks the production fix duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): preserve initially blocked task holds #72030 [closed] fixes — (+45/-0) — n/a: Applies the same lifecycle-consistent event repair with narrower regression coverage; closed explicitly as a duplicate of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• [verified] fix(kanban): preserve initial blocked state #74839 [closed] fixes — (+170/-31) — n/a: Combines the duplicate event insertion with independent promote_task() transaction hardening; closed after contributor triage required splitting the promotion work from fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735.
• fix(kanban): keep initial_status='blocked' tasks sticky #76718 [closed] fixes — (+118/-0) — n/a: Converged to the narrow event repair after dropping reviewed block_kind and reclaim conflation, then closed in favor of fix(kanban): emit blocked event when initial_status='blocked' so sticky-block guard fires #34735 as a recent minimal reference.
• feat(kanban): add audited update-body command #76765 related — (+203/-0) — n/a: Adds an audited update-body command and does not repair the initial-block promotion cause; its keep_open review requests complete hash, length, and no-body-text payload assertions.

Duplicates

#39085, #41497, #46274, #46565, #48437, #61890, #71861, #71977, #72030, and the final narrow #76718 substantially duplicate #34735’s create-time blocked event; #39832, #48253, and #64830 instead duplicate the intent through created-payload inference, while #35832, #66993, #69679, #70733, #71147, #69531, and #74839 add separable broader behavior.

Suggested consolidation

Keep #34735 open with its recorded salvage path: transplant the narrow post-created event insertion and lifecycle tests onto current main. Despite keep_open reviews on #39085, #41497, #46565, #64830, #71861, and #71977, close them as duplicates of #34735 because their diffs show either the same event repair or the less complete payload-inference approach; author action on #69531 is to rebase and reconcile its broader recorded best fix with current main, while #69679 must initialize recurrence to 0, #34756/#35832/#66993/#70733/#71147 should split salvageable extras, and #76765 should remain separate pending the requested audit/privacy assertions.

Complex graph

flowchart TD
    classDef open fill:#dbeafe,stroke:#1d4ed8,color:#1e3a8a
    classDef merged fill:#dcfce7,stroke:#15803d,color:#14532d
    classDef closed fill:#e5e7eb,stroke:#6b7280,color:#1f2937
    classDef unverified fill:#f3f4f6,stroke:#9ca3af,color:#374151
    classDef best stroke-width:3px,stroke:#b45309
    classDef target stroke-width:3px,stroke:#4338ca
    I39609(["issue #39609 (open)"])
    I47777(["issue #47777 (open)"])
    subgraph Dup34735 ["PRs duplicating each other"]
        P34735["PR #34735 (open)"]
        P35832["PR #35832 (open)"]
        P39085["PR #39085 (open)"]
        P39832["PR #39832 (closed)"]
        P41497["PR #41497 (open)"]
        P43479["PR #43479 (closed)"]
        P46274["PR #46274 (closed)"]
        P46565["PR #46565 (open)"]
        P48253["PR #48253 (closed)"]
        P48437["PR #48437 (closed)"]
        P60100["PR #60100 (closed)"]
        P61890["PR #61890 (closed)"]
        P64830["PR #64830 (open)"]
        P70733["PR #70733 (open)"]
        P71147["PR #71147 (open)"]
        P71861["PR #71861 (open)"]
        P71977["PR #71977 (open)"]
        P72030["PR #72030 (closed)"]
        P74839["PR #74839 (closed)"]
        P76718["PR #76718 (closed)"]
    end
    P64830 -.->|partial| I39609
    P64830 -.->|partial| I47777
    class I39609 open
    class I47777 open
    class P34735 open
    class P35832 open
    class P39085 open
    class P39832 closed
    class P41497 open
    class P43479 closed
    class P46274 closed
    class P46565 open
    class P48253 closed
    class P48437 closed
    class P60100 closed
    class P61890 closed
    class P64830 open
    class P70733 open
    class P71147 open
    class P71861 open
    class P71977 open
    class P72030 closed
    class P74839 closed
    class P76718 closed
    class P34735 best
    class P34735 best
    class P64830 target
    click I39609 "https://github.com/NousResearch/hermes-agent/issues/39609"
    click I47777 "https://github.com/NousResearch/hermes-agent/issues/47777"
    click P34735 "https://github.com/NousResearch/hermes-agent/pull/34735"
    click P35832 "https://github.com/NousResearch/hermes-agent/pull/35832"
    click P39085 "https://github.com/NousResearch/hermes-agent/pull/39085"
    click P39832 "https://github.com/NousResearch/hermes-agent/pull/39832"
    click P41497 "https://github.com/NousResearch/hermes-agent/pull/41497"
    click P43479 "https://github.com/NousResearch/hermes-agent/pull/43479"
    click P46274 "https://github.com/NousResearch/hermes-agent/pull/46274"
    click P46565 "https://github.com/NousResearch/hermes-agent/pull/46565"
    click P48253 "https://github.com/NousResearch/hermes-agent/pull/48253"
    click P48437 "https://github.com/NousResearch/hermes-agent/pull/48437"
    click P60100 "https://github.com/NousResearch/hermes-agent/pull/60100"
    click P61890 "https://github.com/NousResearch/hermes-agent/pull/61890"
    click P64830 "https://github.com/NousResearch/hermes-agent/pull/64830"
    click P70733 "https://github.com/NousResearch/hermes-agent/pull/70733"
    click P71147 "https://github.com/NousResearch/hermes-agent/pull/71147"
    click P71861 "https://github.com/NousResearch/hermes-agent/pull/71861"
    click P71977 "https://github.com/NousResearch/hermes-agent/pull/71977"
    click P72030 "https://github.com/NousResearch/hermes-agent/pull/72030"
    click P74839 "https://github.com/NousResearch/hermes-agent/pull/74839"
    click P76718 "https://github.com/NousResearch/hermes-agent/pull/76718"

Loading

Graph: solid arrow = fixes / best fix, dashed arrow = partial or unverified (see edge label); boxed group = PRs duplicating each other; amber border = best fix; indigo border = target; gray node = closed (state tag in the node label).

Cross-PR triage: Reviewed 28 pull requests and 2 issues in this complex. Each diff was read against this issue; Assessment working set: 505 kB of PR diffs, 53 kB of issue/PR text, 35 kB of discussion (55 comments), 74 verify verdicts. verdicts reflect diff content, not PR titles. Part of an automated triage batch.