Best possible solution:

Persist terminal skip bookkeeping for malformed maintainer decisions, then land the packet contract once maintainers accept the generated-state schema and latest checks pass.

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

Yes, source-reproducible for the PR finding: the added malformed maintainer_decision branch records a kept-open result and continues before the normal skip helper stamps apply_checked_at. I did not run tests because this review is read-only.

Is this the best way to solve the issue?

No, not quite; the feature shape is coherent, but malformed intent should still persist terminal skip bookkeeping so apply scans do not keep reprocessing the same bad report.

Full review comments:

• [P2] Persist malformed decision skips — src/clawsweeper.ts:18050-18054
  This new malformed maintainer_decision branch records a kept_open result and then continues without calling markApplyChecked or writeReportMarkdown. Apply windows are ordered by apply_checked_at, so the same malformed generated report can stay at the front of later automatic scans and consume a processed slot every run; route this skip through the normal checked-state bookkeeping while still avoiding GitHub calls.
  Confidence: 0.86

Overall correctness: patch is incorrect
Overall confidence: 0.86

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against def41b3f889a.

Label changes:

• add rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🐚 platinum hermit and patch quality is 🦐 gold shrimp. Replaced prior rating: 🐚 platinum hermit.
• remove rating: 🐚 platinum hermit: Current PR rating is rating: 🦐 gold shrimp, so this older rating label is no longer current.

Label justifications:

• P2: This is a normal-priority automation feature with a bounded but real apply-queue bookkeeping defect.
• merge-risk: 🚨 automation: The diff changes review parsing, report rendering, apply/replay/reconcile state writes, and durable review comments.
• merge-risk: 🚨 compatibility: The diff creates a new durable packet file contract and report frontmatter pointers consumed by companion maintainer tooling.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🐚 platinum hermit and patch quality is 🦐 gold shrimp. Replaced prior rating: 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The provided PR context includes a built-CLI apply-decisions run against temp records with fake gh, showing packet creation, frontmatter pointers, stale-packet deletion, and pointer clearing.
• proof: sufficient: Contributor real behavior proof is sufficient. The provided PR context includes a built-CLI apply-decisions run against temp records with fake gh, showing packet creation, frontmatter pointers, stale-packet deletion, and pointer clearing.

Acceptance criteria:

• [P1] node --test --test-name-pattern "malformed maintainer decisions|decision packet" test/clawsweeper.test.ts test/decision-packets.test.ts.
• [P1] pnpm run check.

What I checked:

• Repository policy read: AGENTS.md was read fully; its generated-state layout, apply-lane safety, and release-owned changelog guidance are relevant to this PR. (AGENTS.md:1, def41b3f889a)
• Current main capability check: Current main has no decision-packet implementation or decision_packet frontmatter support; the only maintainerDecision match is an unrelated repair workflow helper. (def41b3f889a)
• Decision packet persistence: The PR adds syncDecisionPacketRecord, which writes packet JSON, removes stale packet files, and updates decision_packet_path and decision_packet_sha256 frontmatter. (src/decision-packets.ts:267, 927668a34ea3)
• Blocking apply bookkeeping defect: The added malformed maintainer_decision catch branch records a kept_open result and continues without calling the helper that stamps and persists apply_checked_at. (src/clawsweeper.ts:18050, 927668a34ea3)
• Test gap: The malformed decision regression verifies no GitHub mutation and the apply result, but it does not assert that the report gains apply_checked_at. (test/clawsweeper.test.ts:435, 927668a34ea3)
• Live PR state: GitHub reports the PR head as mergeable at 927668a; pnpm check was still in progress at the final live check while CodeQL jobs had passed. (927668a34ea3)

Likely related people:

• steipete: Authored the latest PR commits that make decisions Codex-authored and has recent current-main apply/review policy work in the same area. (role: likely follow-up owner; confidence: high; commits: 0432e4d8e8bf, cdcb851924a7, f4bc90093196; files: src/clawsweeper.ts, prompts/review-item.md, VISION.md)
• brokemac79: Has merged adjacent apply, review, state-scan, and dashboard automation work, including apply queue rotation that depends on apply_checked_at behavior. (role: recent area contributor; confidence: high; commits: 5bdf6f31c904, 542a35738928, 44f0d7ac1222; files: src/clawsweeper.ts, test/clawsweeper.test.ts, README.md)
• vincentkoc: Git history ties the central apply/review report parsing and queue foundations to Vincent Koc, with recent current-main label reconciliation work nearby. (role: introduced behavior; confidence: medium; commits: ecbfc4fbfecb, 0d6aa5054c5a; files: src/clawsweeper.ts, .github/workflows/sweep.yml)
• RomneyDa: The live PR is assigned to RomneyDa, and current-main history shows Dallin Romney split and maintained nearby apply/review test coverage. (role: assigned reviewer and adjacent test owner; confidence: medium; commits: 323f7bdf7236, 1c19987cde97, 8d8565bdea3b; files: test/clawsweeper.test.ts, test/apply-same-author-pair-close.test.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.

• reviewed 2026-07-03T17:05:44.291Z sha ce6591d :: needs maintainer review before merge. :: none