Codex review: needs real behavior proof before merge.

Workflow note: Future ClawSweeper reviews update this same comment in place.

Summary
The PR detaches accepted queued user-request followups from the source abort signal while preserving room-event abort metadata and adds focused regression coverage.

Reproducibility: yes. at source level: current main attaches opts.abortSignal to queued user followups and the queued runner reuses that source signal when queued.abortSignal is absent. I did not run tests because this review is read-only, but the PR supplies focused before/after regression output for the same path.

PR rating
Overall: 🦐 gold shrimp
Proof: 🦐 gold shrimp
Patch quality: 🦞 diamond lobster
Summary: The patch is focused and well tested at source level, but merge readiness is capped by insufficient real Telegram behavior proof for the exact queue/supersede race.

Rank-up moves:

• Add redacted Telegram topic logs or a short recording showing a queued user request accepted while another turn is active and then draining after a later superseding turn.
• Include diagnostics that show the queue/supersede/drain path while redacting chat IDs, phone numbers, tokens, non-public endpoints, and message contents.

PR egg
🎁 Pass real behavior proof to wake the egg and unlock a hatchable treat.

Real behavior proof
Needs stronger real behavior proof before merge: The PR includes redacted patched-runtime Telegram logs, but they do not demonstrate the exact queued-after-supersede path; add redacted logs, terminal output, or a recording that shows that path and update the PR body for re-review.

Mantis proof suggestion
A native Telegram proof can directly show whether the queued topic user message is preserved and answered after a later superseding turn. A maintainer can ask Mantis to capture proof by posting a new PR comment that starts with the OpenClaw Mantis account mention, followed by:

telegram desktop proof: verify that a queued Telegram topic user message drains after a later superseding turn, with room-event cancellation still aborting ambient work.

Risk before merge
Why this matters: - The live proof currently shows same-topic inbound and outbound messages from a patched gateway, but not an accepted queued user turn draining after a later superseding turn.

• The change affects generic queued CLI and embedded followup abort propagation, so maintainers should verify both user-request survivability and room-event cancellation before merge.

Maintainer options:

1. Require exact Telegram queue proof (recommended)
   Ask for redacted logs, terminal output, or a short recording from the patched branch showing a queued Telegram topic user request accepted while another turn is active and then drained after a later superseding turn.
2. Accept source and focused-test proof
   Maintainers can choose to merge on the source-level fix and before/after regression tests while explicitly accepting that the exact live Telegram race was not captured.
3. Pause until Mantis or live lane is available
   If exact proof cannot be captured from the contributor setup, pause the PR until Mantis, Crabbox, or another live Telegram lane can exercise the queue/supersede path.

Next step before merge
The remaining action is exact live Telegram proof and maintainer review, not a safe automated code repair.

Security
Cleared: The diff only changes in-process abort propagation and focused tests; it does not touch secrets, dependencies, CI, install scripts, package metadata, or external code execution paths.

Rebased the draft branch onto current origin/main after opening.

• Head SHA: 56b6731d2290c19413863c52e5adf73bc8a0f4b1
• Change after rebase: no code changes beyond replaying the same queued-followup fix onto latest main
• Focused validation rerun: node scripts/run-vitest.mjs src/auto-reply/reply/get-reply-run.media-only.test.ts src/auto-reply/reply/followup-runner.test.ts src/auto-reply/reply/queue.collect.test.ts extensions/telegram/src/bot-message-dispatch.test.ts -> Test Files 4 passed (4), Tests 220 passed (220)

Refreshed the draft branch head after the proof-body update so the PR gets a clean check set.

• Head SHA: 695478dc167f6cd54cfe318fe037837494bf832a
• Code tree change from prior head: none; commit was amended only to refresh SHA after updating the PR body
• Focused validation rerun: node scripts/run-vitest.mjs src/auto-reply/reply/get-reply-run.media-only.test.ts src/auto-reply/reply/followup-runner.test.ts src/auto-reply/reply/queue.collect.test.ts extensions/telegram/src/bot-message-dispatch.test.ts -> Test Files 4 passed (4), Tests 220 passed (220)

Queue-mode and before/after proof for the queued-followup fix.

Queue mode is still respected

• src/auto-reply/reply/get-reply-run.ts resolves resolvedQueue.mode first, derives activeRunQueueMode, and calls resolveActiveRunQueueAction(...) before a busy turn can be enqueued.
• src/auto-reply/reply/agent-runner.ts only calls enqueueFollowupRun(...) when activeRunQueueAction === "enqueue-followup", and it passes the original resolvedQueue into enqueueFollowupRun(...) unchanged.
• This patch only changes the abort metadata carried by an already accepted queued followup. It does not make interrupt, drop, steer, followup, or collect choose a different queue action.

Before proof

Setup: checked out parent d761b98adc6ff2c824b95d0b2a7a54dfa1177d18 in a separate temp worktree, applied only the new regression tests, and ran:

node scripts/run-vitest.mjs src/auto-reply/reply/get-reply-run.media-only.test.ts src/auto-reply/reply/followup-runner.test.ts

Result on old code:

Test Files  2 failed (2)
Tests  2 failed | 110 passed (112)

FAIL ... followup-runner.test.ts > does not inherit source abort signals for queued user followups
AssertionError: expected AbortSignal { aborted: true } to be undefined

FAIL ... get-reply-run.media-only.test.ts > detaches queued user requests from superseded source abort signals
AssertionError: expected AbortSignal { aborted: false } to be undefined

After proof

Ran the same command on PR head 695478dc167f6cd54cfe318fe037837494bf832a:

Test Files  2 passed (2)
Tests  112 passed (112)

This proves the pre-patch behavior both attached the source abort signal to queued user requests and re-inherited an already-aborted source signal during queued execution; the PR head removes both while leaving the queue-mode decision path intact.

@clawsweeper re-review

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event repository_dispatch).
Result: the existing ClawSweeper review comment will be edited in place when the review finishes.

Re-review progress:

• State: Complete
• Detail: The targeted re-review finished, the durable review comment was updated, and the synced verdict was routed.
• Run: https://github.com/openclaw/clawsweeper/actions/runs/26067558140
• Updated: 2026-05-19T00:06:00.230Z

Updated the PR body with redacted current live Telegram topic proof from the active patched gateway (topic:8185 inbound messages followed by outbound threadId=8185 replies) and noted that the Mantis proof lane was blocked by a conflicted Telegram bot connection.

@clawsweeper re-review

@clawsweeper re-review

Superseded by #83827. Refiled with the RCA, before/after regression proof, queue-mode proof, live Telegram-topic log proof, blocked Mantis proof, and the later topic 22684 client-close classification consolidated in the new PR body.