Best possible solution:

Land only after removing the changelog edit, aligning schema help/type docs with the recurring retry contract, and adding redacted real behavior proof or a maintainer proof override for the cron retry path.

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

Yes, source-reproducible: current main applies transient retry policy only to one-shot jobs while recurring errors use normal error backoff. I did not run tests because this was a read-only review.

Is this the best way to solve the issue?

No, not merge-ready as-is: the runtime approach is plausible, but the release-owned changelog edit, stale config help/type docs, and missing real behavior proof need resolution first.

Full review comments:

• [P2] Remove the hand-written changelog entry — CHANGELOG.md:5-9
  This normal bug-fix PR adds an Unreleased block, but the repo rules say CHANGELOG.md is release-owned and normal PR fixes should carry release-note context in the PR body, squash message, or direct commit context. Leaving this entry mixes hand-written content into generated release notes, so remove the CHANGELOG.md change.
  Confidence: 0.95
• [P2] Align cron retry help text with recurring retries — docs/gateway/configuration-reference.md:1275
  This doc change broadens cron.retry to recurring jobs, but src/config/schema.help.ts and src/config/types.cron.ts still describe the same config as one-shot-only. Users consulting config help or exported config types would get the old contract, so update those surfaces with the recurring retry semantics before merge.
  Confidence: 0.88

Overall correctness: patch is incorrect
Overall confidence: 0.88

AGENTS.md: found and applied where relevant.

Codex review notes: model gpt-5.5, reasoning high; reviewed against e7fb8cabb681.

Label changes:

• add P2: This is a bounded cron reliability improvement with blocking review items but limited blast radius.
• add merge-risk: 🚨 compatibility: Merging would make existing cron.retry settings affect recurring jobs, changing upgraded users' retry cadence and model-call behavior.
• add rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• add status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body lists tests and checks only; before merge, add redacted live output, logs, terminal screenshot, recording, or artifact showing a recurring cron transient retry after the patch, then update the PR body to trigger review or ask a maintainer to comment @clawsweeper re-review.

Label justifications:

• P2: This is a bounded cron reliability improvement with blocking review items but limited blast radius.
• merge-risk: 🚨 compatibility: Merging would make existing cron.retry settings affect recurring jobs, changing upgraded users' retry cadence and model-call behavior.
• rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body lists tests and checks only; before merge, add redacted live output, logs, terminal screenshot, recording, or artifact showing a recurring cron transient retry after the patch, then update the PR body to trigger review or ask a maintainer to comment @clawsweeper re-review.

PR surface:

Source +95, Tests +178, Docs +6, Other -3. Total +276 across 6 files.

What I checked:

• Repository policy read: The full root AGENTS.md was read; it marks CHANGELOG.md as release-owned and says normal fixes should put release-note context in the PR body, squash message, or commit instead. (AGENTS.md:192, e7fb8cabb681)
• Scoped guides read: The relevant docs and scripts scoped AGENTS.md files were read; docs changes are governed by the docs guide and script changes by the scripts guide. Public docs: docs/AGENTS.md. (docs/AGENTS.md:1, e7fb8cabb681)
• Current main behavior: Current main still documents cron retry as one-shot-only and its recurring error branch applies generic error backoff instead of the transient retry policy. (src/cron/service/timer.ts:1006, e7fb8cabb681)
• PR runtime change: The PR adds a shared retry decision helper and uses it in the recurring job error branch to schedule a transient retry before the normal next scheduled run when the retry backoff lands first. (src/cron/service/timer.ts:1052, 89fbe0dc7239)
• Regression coverage: The branch adds tests for recurring transient retry, exhausted recurring retries falling back to the normal schedule, and preserving every-cadence after retry success. (src/cron/service/timer.regression.test.ts:478, 89fbe0dc7239)
• Changelog edit: The PR adds an Unreleased changelog entry despite the release-owned changelog policy; an existing Codex review comment also flagged this as a P2 issue. (CHANGELOG.md:5, 89fbe0dc7239)

Likely related people:

• steipete: Recent current-main history shows Peter Steinberger carrying the cron timer/config documentation import and recent cron delivery ergonomics work; he also authored this PR branch. (role: recent area contributor; confidence: high; commits: d5bbf3033c9f, 2209faef405f, cd0bc5195389; files: src/cron/service/timer.ts, docs/gateway/configuration-reference.md, src/config/schema.help.ts)
• samzong: Recent current-main history includes adjacent cron schedule identity work in a touched file, which may help review schedule identity side effects. (role: adjacent contributor; confidence: medium; commits: c237de552ac3; files: src/cron/schedule-identity.ts)

• 🦀 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.

• 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.