[kyle-sexton]

No related issue: review-input hardening from #9875 review (F1, F5)

Summary

This re-pins the code and security review callers to ci-workflows v0.65.9 (4fbceb4b75ffd568dfa5db31de0d0e753c0c64e6). That release is ci-workflows#884, which closes F5 from the #9875 review.

Before v0.65.9, the review's full diff was a plain git diff in the head checkout. A file could reach the reviewer as Binary files differ in three ways:

• a .gitattributes marks it -diff, including this repository's own -diff on package-lock.json, packages.lock.json and *.lock;
• a .gitattributes marks it binary;
• the file contains a NUL byte.

v0.65.9 writes the diff with --text --no-textconv. F1, head instruction files loading into the reviewer, did not reproduce. The evidence is in ci-workflows#884.

Fix

• pr-review-hosted.yml and pr-review-security-hosted.yml call pr-review-cli.yml@4fbceb4b75ffd568dfa5db31de0d0e753c0c64e6 # v0.65.9.
• scripts/ci-review-lanes.test.sh expects the new pin.
• Inputs, secrets and permissions are unchanged.
• The v0.65.9 runner-policy contract arrived through the standards sync chore: sync standards components #9908 (standards#878).

Verification

• scripts/ci-review-lanes.test.sh: PASS=88 FAIL=0.
• actionlint is clean on both callers.
• zizmor --offline reports no findings on both callers.
• This pull request's own review lanes run on v0.65.9, because both callers run on pull_request.

Related

• ci-workflows#884 (v0.65.9) and ci-workflows#886 (lock-file fixture, test only).
• standards#878, chore: sync standards components #9908.
• chore(ci): re-pin the review callers to ci-workflows v0.65.8 #9848 (the v0.65.8 re-pin this follows).

🤖 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 2026-10-25T01:13:48.696260Z	adcb8ea	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-security-review reviewed adcb8eac3eabbff8764c9b43ca3934eead3bf5b7: 0 finding(s), 0 as inline comments.

No security issues found in PR #9910 at head adcb8ea.

The diff changes three files. Two callers, pr-review-hosted.yml and pr-review-security-hosted.yml, move their pr-review-cli.yml pin from v0.65.8 (4f79cc92…) to v0.65.9 (4fbceb4b…). scripts/ci-review-lanes.test.sh is updated to expect the new SHA and version.

• Pin form: both callers still use a full 40-character commit SHA with a version comment, from the same org's ci-workflows repo. Nothing is unpinned or loosened.
• Permissions and inputs: contents: read and pull-requests: write are unchanged. publish: comments, hold-label: needs-human, lane: and the secrets are untouched. The test script still checks all of these.
• Test script: it still verifies the job id, publish: comments, hold-label and the exact pin, so a drifted or unpinned reference still fails the test.
• Pin target not checked: this run has no network access, so I could not confirm that 4fbceb4b… is the v0.65.9 tag commit. The PR cites ci-workflows#884 for the change, and the SHA-pin and zizmor checks cover the pin form. A maintainer who wants that assurance can compare the SHA with the v0.65.9 release tag.
• Trusted paths: none of the three changed files match the listed trusted-path globs.
• PR context: it contained no embedded instructions.
• Manifests and instruction surfaces: the diff has no plugin manifest changes and no instruction-surface deletions.
• Spend: the diff only changes which workflow version runs. It adds no paid-plan or billing change, and I am not treating it as a cost hold.

critical-findings: 0

[github-actions]

claude-review reviewed adcb8eac3eabbff8764c9b43ca3934eead3bf5b7: 0 finding(s), 0 as inline comments.

No findings. This PR re-pins the code and security review callers from ci-workflows v0.65.8 to v0.65.9.

• Pin consistency: both callers (.github/workflows/pr-review-hosted.yml:171 and .github/workflows/pr-review-security-hosted.yml:176) use 4fbceb4b75ffd568dfa5db31de0d0e753c0c64e6 # v0.65.9. scripts/ci-review-lanes.test.sh expects the same SHA and version comment in its grep, its ok message and its fail message.
• Runner policy: .github/standards/runner-policy/policy.json:1098 already has an entry for the new SHA, which matches the description's note that it arrived through the standards sync (chore: sync standards components #9908). I did not read that entry's contents.
• Leftover old pins: the old SHA remains only as a historical entry in policy.json. The v0.65.8 links in plugins/review/skills/*/SKILL.md point at a specific release's source, so they are not stale pins.
• Scope of the change: inputs, secrets and permissions are unchanged, as the description says.
• Trusted paths: none of the three changed files matches a trusted-path glob.
• Closing issue: none.
• Not verified: I had no shell. The author's reported results (test script PASS=88, actionlint and zizmor clean) are author-claimed, and I did not run them. I did not read the v0.65.9 release itself, so I cannot confirm what it contains.
• Cost: this change adds no billing. It changes which workflow version the review lanes run.