[kyle-sexton]

No related issue: found while verifying #8994's probe on ci-workflows#972; no open issue owns it.

Summary

The ready reference and the review-venue reference said every draft head reads local (tests absent). That is true only for a head pushed while the PR is a draft. A PR moved back to draft keeps the test-lane results of a head pushed while it was ready, so its venue reads ci healthy. The probe on ci-workflows#972 (comment on #8994) showed this.

Fix

• plugins/source-control/skills/pull-request/reference/ready-for-review.md: only a head pushed while the PR is a draft reads local; a PR moved back to draft keeps its ready head's results.
• plugins/source-control/skills/pull-request/reference/review-venue.md: the same correction on the tests local absent line.
• .changes/source-control/ patch fragment.

Verification

• bash scripts/check-changelog-fragments.sh --check-required origin/main: "Every fragment-mode plugin this change set changes (1) has a fragment".
• The probe's venue line on ci-workflows#972's re-drafted head read tests ci healthy: 1 test-lane check runs succeeded on e50d5598... (comment).

Related

• source-control: observe a ready run that skips the local verify #8994 (closed; the probe that found this)
• fix(pr-review-cli): read only a reviewed-head marker from the same caller ci-workflows#972

🤖 Generated with Claude Code

https://claude.ai/code/session_01VcupomFZgq3z3EzLHjxPB9

[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 2026-11-03T06:24:29.522936Z	2a6e5ad	Draft marked ready

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 65768849841d377bf6997ba8b1f65ec687af3b8b: read 3 of 3 changed file(s); 0 finding(s), 0 as inline comments.

Code review: no findings

This PR changes documentation only. It corrects two reference docs for /source-control:pull-request and adds a changelog fragment:

• reference/ready-for-review.md (§2.5.3): the doc used to say every draft head reads local. It now says that only a head pushed while the PR is a draft reads local. A PR moved back to draft keeps the results of a head pushed while it was ready.
• reference/review-venue.md: makes the same correction on the tests local absent line.
• .changes/source-control/fix-draft-head-venue-wording-8994a1c2.md: a bump: patch fragment under ### Fixed. A patch bump fits a docs correction.

Checks done

• I read every hunk in all 3 changed files. The two docs now say the same thing. The wording also matches how the venue reader works: it reads the check runs on the head commit, not whether the PR is currently a draft.
• Path rules: .claude/rules/skill-bodies-state-current-rules.md and .claude/rules/cost-claims.md both match the changed skill reference files. Neither rule is broken. The new sentences describe how this repository's own test lanes behave, and they state no price or cost figure.
• No plugin manifest changed, so the manifest checks don't apply.
• The PR closes no issue, so there are no acceptance criteria to check.
• Prior findings: the only earlier post was a no-findings summary at head 2a6e5ade, so nothing is repeated. This is round 0 of 3; the cap doesn't apply.
• This head is a merge from main. The PR's diff is the same as at the head reviewed before.

Not verified: the check-changelog-fragments.sh run and the probe's venue reading on ci-workflows#972 are claims from the PR author. I didn't run them myself.

[github-actions]

claude-security-review reviewed 65768849841d377bf6997ba8b1f65ec687af3b8b: read 3 of 3 changed file(s); 0 finding(s), 0 as inline comments.

Security review: no findings

I found no security issues. I read every change in all 3 changed files, so this review is complete.

What changed. The pull request edits three documentation files:

• .changes/source-control/fix-draft-head-venue-wording-8994a1c2.md: a new patch changelog fragment.
• plugins/source-control/skills/pull-request/reference/ready-for-review.md: says only a head pushed while the PR is a draft reads local. A PR moved back to draft keeps the test-lane results of a head pushed while it was ready.
• plugins/source-control/skills/pull-request/reference/review-venue.md: makes the same correction on the tests local absent line.

Instruction-surface check. Both reference files belong to a skill body, so I checked whether the edit weakens a standing rule. It does not:

• The edited sentences describe what the venue script reports. They set no rule of their own.
• The gate's decision rule is unchanged. It still runs in full when the line reads local, when the venue cannot be read, or when the run passed --verify-local.
• A re-drafted head whose venue reads ci healthy skips suites only because CI test lanes already passed on that same head.
• The line "ci-status gates the merge either way" is still in place, unchanged.

Not affected:

• No trust control, token scope, kill switch, merge-hold script or workflow is touched.
• No plugin manifest is changed.
• No code is changed, and no changed file matches a trusted-path glob.

Prior findings. The earlier round at 2a6e5ade had no findings. The same three files carry the same content at this head, so nothing needed to be reposted.

PR description. It contains no text that tries to steer this review.

critical-findings: 0