Best possible solution:

Keep the narrow test-alignment PR open for normal maintainer readiness and merge handling once it is no longer draft and required checks are acceptable.

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

Not applicable: this is a test-alignment PR, not a bug report with a current-main failure path. Source inspection shows the assertions target existing contracts rather than a new runtime behavior.

Is this the best way to solve the issue?

Yes: updating the narrow tests is the best fix shape because the production prompt, gateway handoff, and tool-card rendering contracts already exist on current main. No broader refactor or compatibility path is needed for this diff.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against cd25de98b1b9.

Label changes:

• add P3: This is a low-risk maintainer test-alignment PR with no runtime, config, migration, or user-facing behavior change.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: The external contributor proof gate is not applicable to this maintainer-labeled MEMBER draft; the PR body still lists focused Testbox, gateway, UI, Codex/ACP, and check:changed validation.

Label justifications:

• P3: This is a low-risk maintainer test-alignment PR with no runtime, config, migration, or user-facing behavior change.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: The external contributor proof gate is not applicable to this maintainer-labeled MEMBER draft; the PR body still lists focused Testbox, gateway, UI, Codex/ACP, and check:changed validation.

PR surface:

Tests +2. Total +2 across 3 files.

What I checked:

• Live PR state: Live GitHub data shows this PR is open, draft, maintainer-labeled, and changes only three test files at head 117ecde. (117ecde2bee7)
• Skill Workshop prompt contract: Current main builds the Skill Workshop prompt as the compact routing and pending-proposal contract that the PR now asserts. (src/agents/skill-workshop-prompt.ts:8, cd25de98b1b9)
• Codex adapter caller: The Codex app-server developer-instruction builder includes the shared Skill Workshop prompt only when a dynamic tool named skill_workshop is available, so the updated Codex test targets the right caller boundary. (extensions/codex/src/app-server/thread-lifecycle.ts:1696, cd25de98b1b9)
• PR Skill Workshop assertion: The PR head checks the stable compact prompt fragments instead of the old long prose removed from the shared contract. (extensions/codex/src/app-server/thread-lifecycle.test.ts:291, 117ecde2bee7)
• Gateway startup handoff contract: Current main waits on deferAccountStartUntil before measuring start-account-handoff, and exits before that handoff when the abort signal or manual stop wins. (src/gateway/server-channels.ts:637, cd25de98b1b9)
• PR gateway test assertion: The PR uses the production deferStartupAccountStartsUntil hook in the test helper and asserts that stop prevents both startAccount and start-account-handoff measurement. (src/gateway/server-channels.test.ts:1254, 117ecde2bee7)

Likely related people:

• vincentkoc: Current-main blame and git log tie the central prompt, gateway handoff, and tool-card renderer lines to 17777b3 and nearby refactor work authored by this contributor. (role: recent area contributor; confidence: high; commits: 17777b3a9ff1, 6418e196b123; files: src/agents/skill-workshop-prompt.ts, src/gateway/server-channels.ts, ui/src/pages/chat/components/chat-tool-cards.ts)
• steipete: Merged PR perf(agents): slim Skill Workshop prompt section to its routing contract #100481 intentionally slimmed the Skill Workshop prompt section to the routing contract that this PR now asserts. (role: feature owner for compact prompt contract; confidence: medium; commits: 47ce4522b6ed; files: src/agents/skill-workshop-prompt.ts, src/agents/tools/skill-workshop-tool.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.