• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

Best possible solution:

Land the worker-liveness watchdog with the focused regression coverage once maintainers accept the session-level proof or obtain a live Telegram smoke for the transport path.

Do we have a high-confidence way to reproduce the issue?

Yes. Current main clearly passes pollingStallThresholdMs while defaulting isolated ingress on, but the threshold is only consumed by the non-isolated polling watchdog path.

Is this the best way to solve the issue?

Yes. Reusing the existing liveness tracker for isolated worker messages is a narrow fix for the documented config contract and avoids adding a new opt-out setting or changing the public config surface.

Label changes:

• add P1: The PR fixes a default Telegram polling stall path that can wedge inbound updates for real channel users.
• add merge-risk: 🚨 message-delivery: The diff changes when the Telegram isolated ingress worker is stopped and restarted in the inbound polling path.
• add merge-risk: 🚨 availability: A watchdog false positive could churn the Telegram isolated polling loop and temporarily reduce channel availability.
• add proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes after-fix AWS Crabbox live output from a runtime TelegramPollingSession harness showing the silent worker stopped and restarted, with the caveat that it did not use a real Telegram bot token/network stall.
• add rating: 🐚 platinum hermit: Current PR rating is 🐚 platinum hermit because proof is 🐚 platinum hermit, patch quality is 🐚 platinum hermit, and Focused implementation and regression coverage with credible session-level proof; live Telegram proof would raise confidence for the transport-risk edge.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes after-fix AWS Crabbox live output from a runtime TelegramPollingSession harness showing the silent worker stopped and restarted, with the caveat that it did not use a real Telegram bot token/network stall.

Label justifications:

• P1: The PR fixes a default Telegram polling stall path that can wedge inbound updates for real channel users.
• merge-risk: 🚨 message-delivery: The diff changes when the Telegram isolated ingress worker is stopped and restarted in the inbound polling path.
• merge-risk: 🚨 availability: A watchdog false positive could churn the Telegram isolated polling loop and temporarily reduce channel availability.
• rating: 🐚 platinum hermit: Current PR rating is 🐚 platinum hermit because proof is 🐚 platinum hermit, patch quality is 🐚 platinum hermit, and Focused implementation and regression coverage with credible session-level proof; live Telegram proof would raise confidence for the transport-risk edge.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes after-fix AWS Crabbox live output from a runtime TelegramPollingSession harness showing the silent worker stopped and restarted, with the caveat that it did not use a real Telegram bot token/network stall.
• proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes after-fix AWS Crabbox live output from a runtime TelegramPollingSession harness showing the silent worker stopped and restarted, with the caveat that it did not use a real Telegram bot token/network stall.

What I checked:

• current default path makes the bug source-reproducible: Current main passes account.config.pollingStallThresholdMs into TelegramPollingSession while defaulting isolatedIngress.enabled to true, so the configured watchdog value is present on the default path. (extensions/telegram/src/monitor.ts:292, e42726204490)
• current main only applies the stall threshold in non-isolated polling: Current main stores #stallThresholdMs and chooses the isolated cycle by default, but the visible watchdog detectStall call using #stallThresholdMs is in #runPollingCycle rather than #runIsolatedIngressCycle. (extensions/telegram/src/polling-session.ts:280, e42726204490)
• PR wires isolated worker messages into liveness: The head revision creates a TelegramPollingLivenessTracker in #runIsolatedIngressCycle and records poll-start, poll-success, poll-error, and spooled activity from the isolated worker before watchdog evaluation. (extensions/telegram/src/polling-session.ts:708, 84b887d820f3)
• PR adds an isolated ingress watchdog restart: The head revision adds an interval in the isolated cycle that calls liveness.detectStall with this.#stallThresholdMs, marks transport dirty, logs/statuses the stall, stops the worker, and forces the cycle forward after the stop grace timer. (extensions/telegram/src/polling-session.ts:808, 84b887d820f3)
• PR extends liveness tracking for worker activity: The head revision adds lastGetUpdatesActivityAt, noteGetUpdatesSuccessCount, and noteGetUpdatesActivity so spooled updates can keep an in-flight isolated poll from being treated as silent. (extensions/telegram/src/polling-liveness.ts:15, 84b887d820f3)
• regression coverage covers restart and active spooling: The PR adds tests for restarting a stalled isolated ingress worker and for keeping the worker alive when spooled messages show activity before the threshold expires. (extensions/telegram/src/polling-session.test.ts:782, 84b887d820f3)

Likely related people:

• Peter Steinberger: Relevant history shows Peter split Telegram polling/session surfaces and deduped the polling-session harness, which are central to reviewing this PR's implementation and tests. (role: major refactor and test-harness contributor; confidence: medium; commits: 0c0f1e34cba0, 53f90af990a6, a185ca283a74; files: extensions/telegram/src/polling-session.ts, extensions/telegram/src/polling-session.test.ts, extensions/telegram/src/polling-liveness.ts)
• wangchunyue: The closest prior watchdog-specific history changed polling-session.ts and polling-session.test.ts to prevent the Telegram polling watchdog from dropping replies. (role: prior polling watchdog contributor; confidence: medium; commits: dd61171f5b3e; files: extensions/telegram/src/polling-session.ts, extensions/telegram/src/polling-session.test.ts)
• pkuGeo: A prior stalled polling fix rebuilt Telegram transport after stalled polling cycles in the same polling-session code path. (role: stalled polling restart contributor; confidence: medium; commits: e035a0d98c91; files: extensions/telegram/src/polling-session.ts)
• Dallin Romney: Current-main blame for the isolated polling cycle and liveness tracker points to a recent commit touching these files, although the commit subject is broader than this specific Telegram behavior. (role: recent current-file contributor; confidence: low; commits: 447a3643c69b; files: extensions/telegram/src/polling-session.ts, extensions/telegram/src/polling-liveness.ts, extensions/telegram/src/telegram-ingress-worker.ts)

Codex review notes: model gpt-5.5, reasoning high; reviewed against e42726204490.

• Eggs appear after the PR passes real-behavior proof. It is here for vibes, not verdicts: it does not change labels, ratings, merge decisions, or automation.
• The shell reacts to review momentum: open follow-up work warms it up, re-review makes it wobble, and a clean final review lets it hatch.
• Hatchability usually comes from sufficient real-behavior proof, no blocking P0/P1/P2 findings, no security attention needed, and clean correctness. A merged PR is already final, so merge makes the egg hatchable independently.
• The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
• Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.