Skip to content

docs(testing): correct the opus claim and add model pointers - #11678

Merged
kyle-sexton merged 2 commits into
mainfrom
docs/6916-testing-pointers
Oct 31, 2026
Merged

kyle-sexton merged 2 commits into
mainfrom
docs/6916-testing-pointers

Conversation

@kyle-sexton

Copy link
Copy Markdown
Member

Refs: #6916

Summary

Delivers the plugins/testing part of the repo-mode /multi-agent:audit-defaults sweep: one corrected model claim and pointers for the model text that had none.

Fix

  • cleanup: no longer says model: opus keeps the classifier off the session's model, since an opus alias resolves to the session's model when the session is on that family. The cost reason for choosing opus stays, now pointing at sub-agents#choose-a-model.
  • audit and agents/test-judge.md: the model-class mapping and the per-call-model precedence point at sub-agents#choose-a-model, with an as-of date and recheck trigger.

No step changes. No edit cuts a rule, so no eval evidence applies.

Verification

  • The as-of stamps in the diff record a raw read of sub-agents#choose-a-model on 2026-10-31 by the authoring session.
  • Changelog fragment: .changes/testing/fix-6916-model-claims-f07b3d28.md.
  • scripts/check-affected.sh: 13 passed, 2 skipped.
  • Left out: claude plugin eval, because no edit here is a cut.

Related

Remaining

🤖 Generated with Claude Code

An opus alias resolves to the session's model when the session is on
that family, so the classifier dispatch cannot promise otherwise. The
cost reason for opus stays. Add pointers with as-of dates and recheck
triggers to the audit and judge model text.

Refs: #6916

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kyle-sexton
kyle-sexton marked this pull request as ready for review October 31, 2026 19:38
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 31, 2026 •

Copy link
Copy Markdown

Codex Review Summary

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

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-31T19:40:44.620500Z 9f9ec74 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

github-actions Bot commented Oct 31, 2026 •

Copy link
Copy Markdown
Contributor

claude-security-review reviewed 56abe5dcced823038502329ff0e8a5899ad5d13e: 0 finding(s), 0 as inline comments.

No security findings in the incremental change since 9f9ec74.

I read the change in plugins/testing/skills/cleanup/SKILL.md (step 4, Classify). It rewords the pointer sentence about how the opus alias resolves. It now says "for the cases where that alias resolves to the session's own model" and "recheck when that section changes those cases". The instruction itself is unchanged: dispatch a fresh-context general-purpose subagent with model: opus, write the table to <work>/classifier-answer.md, and apply the reading rules. I checked it against the instruction-surface lens. No rule was deleted, narrowed or softened, so the diff does not change what the skill forbids or requires.

  • The file is not on the trusted-path list.
  • No manifest, workflow, token scope, trusted-actor, kill-switch or merge-hold file changed in this increment, so there are no manifest <field> lines.
  • I did not run a refute pass or dispatch verifiers because there were no candidates.
  • I did not review the other three files in the PR this round. They are unchanged since the last review.

critical-findings: 0

@github-actions github-actions Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

claude-review: 1 inline finding(s) on 9f9ec74ca64b977bf6692655a16624873f6cb51a; the summary is in this lane's pull request comment.

claude-lane: claude-review

Comment thread plugins/testing/skills/cleanup/SKILL.md Resolved
@github-actions

github-actions Bot commented Oct 31, 2026 •

Copy link
Copy Markdown
Contributor

claude-review reviewed 56abe5dcced823038502329ff0e8a5899ad5d13e: 0 finding(s), 0 as inline comments.

Incremental review of plugins/testing/skills/cleanup/SKILL.md, covering only the change since 9f9ec74.

The change rewords the parenthetical on model: opus in the Classify step. It now says "for the cases where that alias resolves to the session's own model, see … recheck when that section changes those cases". The old wording had a dangling "when … changes when an alias follows the session's model". The new sentence is grammatical and keeps the as-of date, the recheck trigger and the cost rationale for choosing opus. It changes no step and no rule, so I found no correctness or maintainability problem.

I read the changed lines in the head file. I did not run any checks, and the PR's own verification claims (check-affected.sh results) are author-claimed and unverified by me.

No closing issue states acceptance criteria. The PR only refs #6916. The PR context contained no embedded instructions. No manifest files changed. No cost or critical-security hold applies.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kyle-sexton
kyle-sexton added this pull request to the merge queue Oct 31, 2026
Merged via the queue into main with commit 22d7f15 Oct 31, 2026
42 checks passed
@kyle-sexton
kyle-sexton deleted the docs/6916-testing-pointers branch October 31, 2026 21:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant