[kyle-sexton]

No related issue: follow-up to ci-workflows#876. claude-code-plugins' review callers pin to this release in claude-code-plugins#9589.

Summary

Registers pr-review-cli.yml@f133c6651f866845e18595d26cf59d4043b5b642, which is ci-workflows v0.65.3.

Fix

v0.65.3 changes only apply-review-hold/hold.cjs and its test. git diff --stat e7f935fc f133c665 lists those two files and nothing else. pr-review-cli.yml is byte-identical to v0.65.2, so the contract copies the v0.65.2 entry from #866 forward verbatim.

Verification

• node --test runner-policy.test.mjs in components/runner-policy: 345 pass, 0 fail. This includes the new copy-forward test.

Related

• #866 (v0.65.2) and #864 (v0.65.1)

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-24T05:14:46.290678Z	caade90	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[github-actions]

claude-review reviewed caade90a32e5bf01fbb72336932bdb64254a7cac: 0 finding(s), 0 as inline comments.

No findings.

What I checked

• components/runner-policy/policy.json:972-992 adds the pr-review-cli.yml@f133c6651f866845e18595d26cf59d4043b5b642 contract. I compared it by eye with the v0.65.2 entry at policy.json:951-971. The routing, fixedRunsOn, all 11 allowedInputs and both allowedSecrets mappings are identical. Nothing widens: no new input, no new secret, no runner change.
• The new entry is the only key ending in that SHA, so the test's uniqueness assertion holds.
• The new test at runner-policy.test.mjs:9990-10000 follows the v0.65.2 copy-forward pattern. It deep-compares the new entry with the v0.65.2 entry and asserts exactly one key for the new SHA.
• The entry is placed directly after the v0.65.2 entry, matching the ordering of the earlier entries.
• The PR body states the repo's contract: it has a No related issue: line and the Summary, Fix, Verification and Related sections. There are no closing issues, so there are no acceptance criteria to check.

Not verified

• I could not run node --test because the sandbox required approval for the command. The "345 pass, 0 fail" in the PR body is author-claimed and I did not confirm it.
• I did not independently confirm the claim that pr-review-cli.yml is byte-identical between v0.65.2 and v0.65.3. That would need the ci-workflows repo.

Hold for a person

• No cost or trust-gate class applies. The change adds no model call, schedule or runner, and it touches no trusted-actors, CODEOWNERS or .gitmodules file.

[github-actions]

claude-security-review reviewed caade90a32e5bf01fbb72336932bdb64254a7cac: 0 finding(s), 0 as inline comments.

Security review: PR #867 (head caade90)

No security findings. critical-findings: 0

The PR adds one approvedReusableWorkflowContracts entry in components/runner-policy/policy.json. It allowlists pr-review-cli.yml at the ci-workflows v0.65.3 SHA f133c6651f866845e18595d26cf59d4043b5b642. It also adds a test that the entry equals the v0.65.2 entry.

• Entry matches the v0.65.2 entry. I compared it with the v0.65.2 entry (@e7f935fc…, lines 951–971). routing, fixedRunsOn, allowedInputs and allowedSecrets are identical, and nothing is widened. The secrets are the same two REVIEWS_* mappings, and the runner stays hosted-only on ubuntu-24.04.
• Pinned by full commit SHA. The new key is a full 40-character SHA, not a tag or branch.
• Test. The new test is a deepEqual of the new entry against the v0.65.2 entry. It also checks that exactly one entry exists for the new SHA. It adds no executable surface.
• Not checked: I could not confirm that f133c66… is what the v0.65.3 tag points to. The gh api call needed approval and was not run. The commit message says the workflow file is unchanged from v0.65.2. An owner can confirm the SHA against the tag, but nothing in the diff suggests a mismatch.
• Skip gate and hold. No plugin manifest changed and no instruction was deleted or softened, so the instruction-surface and manifest checks do not apply. The PR touches no trust-gate path (trusted-actors, check-trusted-trigger, CODEOWNERS, .gitmodules).