Skip to content

docs(review): record the bundled code-review scope as of Claude Code 2.1.328 - #18023

Open
melodic-automation-lanes[bot] wants to merge 3 commits into
mainfrom
work-resolve-issue/issue-17908-44938436628-1
Open

melodic-automation-lanes[bot] wants to merge 3 commits into
mainfrom
work-resolve-issue/issue-17908-44938436628-1

Conversation

@melodic-automation-lanes

@melodic-automation-lanes melodic-automation-lanes Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Refs: #17908

Summary

Opened by the work-resolve-issue lane. In bundled-code-review.md's extraction record, the old 2.1.263 menu description, which had the bundled /code-review review for bugs and cleanups, stays as history and is no longer presented as the current scope. Following review feedback, the record does not paraphrase the current scope. It notes that nothing in this lane acts on that scope and points to the /code-review row of the commands reference, where the scope is read live, with the 2.1.328 changelog entry to correlate with.

The second done-when item is still open: re-extracting the two code-review rows in docs/native-surfaces/records.json, then regenerating docs/native-surfaces.md. That needs /harness-ops:audit-native-overlap on a host with the installed Claude Code binary. This lane cannot run inventory.py or overlap.py, and the issue rules out a hand edit, so this PR leaves that item and the issue stays open.

Remaining work

  • Re-extract both code-review rows in docs/native-surfaces/records.json to the current native description through /harness-ops:audit-native-overlap on a host with the installed Claude Code binary, then regenerate docs/native-surfaces.md from the store

Not verified

  • Done-when item 2: re-extract the two code-review rows in docs/native-surfaces/records.json (lines 22 and 2203) and regenerate docs/native-surfaces.md. This needs /harness-ops:audit-native-overlap with an inventory.py extraction of the installed binary, then overlap.py generate. The lane cannot run…
  • The Claude Code version comes from the 2.1.328 changelog entry and the commands reference, not a string search of an installed binary, since this lane cannot run one
  • tidy: skipped in part. Passes 1-3 had nothing to do (markdown only), and passes 4-5 ran by reading the text and made no edits. The ai-slop deterministic detector (detect.sh) did not run because the lane's Bash allows only the named wrappers
  • check-affected skipped typos, comment-hygiene, size-check and validate-plugins (native-overlap registry and others) in this lane; they are left to the PR's CI
  • The review found a stale 'As of 2026-10-11' date on table row 1 of bundled-code-review.md. 2.1.328 names /code-review, so its recheck trigger fired, but the row's claim (bundled, alias review, user-invocable) still matches the live commands row. It was left unchanged as outside this issue
  • Untracked dotfiles at the checkout root (.bashrc, .profile, .gitconfig, .idea, .vscode, .mcp.json and others) were not created by this run and are not part of the change

Fix

  • plugins/review/skills/code-review/reference/bundled-code-review.md: the Extraction record keeps the 2.1.263 observation as history and adds a pointer paragraph. The paragraph says that description is not the current scope and that this lane does not act on the bundled skill's scope. It points to the /code-review row of the commands reference for the live scope, with the 2.1.328 changelog entry to correlate with. It has no as-of date, no recheck trigger and no paraphrase of upstream behavior.
  • .changes/review/work-resolve-issue-issue-17908-44938436628-1-7e2c5e74.md: patch changelog fragment for the review plugin describing the record update

Verification

The runner ran each check from the base copy against d9fdb72:

  • scripts/check-changelog-fragments.sh --check: passed
  • scripts/check-changed-skills.sh --merge-base: passed

The run reports:

  • read https: //raw.githubusercontent.com/anthropics/claude-code/main/CHANGELOG.md: the 2.1.328 section says "Changed /code-review on most models to report only bugs and CLAUDE.md violations; run /simplify for reuse, simplification...", and the newest heading was 2.1.330
  • read https: //code.claude.com/docs/en/commands.md: the /code-review row says it reviews for correctness bugs, and depending on your model and effort level the review also covers cleanup opportunities. Alias /review
  • check-changelog-fragments: all 4 fragments valid (run after the final edit)
  • check-changed-skills: code-review PASS, 0 errors, 5 warnings that already existed in SKILL.md, which this change does not touch
  • check-affected: 14 passed, 9 skipped (tools not installed, network fetch denied, or checks that run code from the checked tree), the rest CI-only
  • fresh-context review subagent: no high findings. It checked the changelog and commands claims and the #21328 anchor. It flagged an unsupported 'before 2.1.328' claim and imprecise fragment wording, and both are fixed
  • The lane run: https://github.com/melodic-software/claude-code-plugins/actions/runs/44938436628

Left out: the items under Not verified.

Related

#17908

Lane-Activity: work-resolve-issue/resolve-issue
@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

claude-security-review reviewed be91a01d4a0e3f0170fab2446dce947feb58bbb1: read 1 of 1 changed file(s); 0 finding(s), 0 as inline comments.

Security review (incremental, blocking only): no security issues

I read the whole incremental diff (87c556c7..be91a01d) for the one listed file:

  • plugins/review/skills/code-review/reference/bundled-code-review.md: the extraction record's pointer paragraph now also records, with a date, the bundled /code-review scope as of Claude Code 2.1.328. Per changelog items 2.1.328-124/125, it reports only bugs and CLAUDE.md violations, sends cleanups to /simplify, and marks each finding Important, Pre-existing or Nit. The paragraph still says nothing in this lane acts on that scope, and it still points to the live commands reference and the changelog.

Security relevance: none.

  • The file matches no trusted-path glob.
  • The change touches no plugin manifest, userConfig key, hook, workflow, script, token or trust control.

Instruction surface: the change adds a descriptive observation about a native command's behaviour. It deletes, narrows or softens no rule that forbids an action, so no protected-class check applies.

Embedded text: nothing in the PR context tells the lane to pass the PR, skip a check or drop a finding.

Prior findings: this lane's only earlier post was a summary with 0 findings. There is nothing to restate.

This review was blocking only, since this lane already reviewed an earlier head.

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 d9fdb72849690f4d610b1172aedac4ea594382b6; the summary is in this lane's pull request comment.

claude-lane: claude-review

@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

claude-review reviewed be91a01d4a0e3f0170fab2446dce947feb58bbb1: read 1 of 1 changed file(s); 1 finding(s), 1 as inline comments.

Code review: PR #18023 (incremental, 87c556c → be91a01)

This push brings back the defect this lane flagged earlier, so there is one blocking finding. This lane already reviewed an earlier head, so this review covered blocking findings only.

What changed since the last review. One file changed: plugins/review/skills/code-review/reference/bundled-code-review.md. In its Extraction record, the head commit adds a sentence to the pointer paragraph at L42-L45. The sentence states the bundled /code-review scope "as of Claude Code 2.1.328, read 2026-11-21". That scope is:

  • it reports only bugs and CLAUDE.md violations on most models;
  • it sends reuse, simplification and efficiency cleanups to /simplify;
  • it marks each finding Important, Pre-existing or Nit.

The owner asked for this in a reply on another thread.

Blocking finding (inline). The added sentence breaks .claude/rules/skill-bodies-state-current-rules.md, the same problem as this lane's earlier finding discussion_r4654539517. That thread is marked lines changed since resolution: yes, so the rule says to raise it again. I've noted the owner's decision in the inline comment. The comment gives two ways to settle it: drop the paraphrase, or record an exception to the rule.

Settled threads. The owner's reply (kyle-sexton) on the thread at L1 is a decision to state the 2.1.328 scope in the record. The finding that thread answers was withheld, so I can't confirm it raised the same point. The rule still forbids this text, so I reported it and left the trade-off to the owner.

Closing issue. The PR says Refs: #17908 and closes no issue, so there are no acceptance criteria to check.

Loosened checks. The prompt gave no loosened_check_verdict: or loosened_check_patterns: line, so I assumed fail and no consumer entries. It gave no loosened_check_decision_record: line either, so I did not check the citation requirement. This push changes prose only and loosens no check.

Other checks.

  • scripts/check-as-of-recheck.sh will not catch this paragraph. Its stamp pattern needs as of YYYY-MM-DD, and the text reads "as of Claude Code 2.1.328, read 2026-11-21". I found this by reading the script; I did not run it.
  • No plugin manifest changed, so there are no manifest lines.
  • No paved-path entry applies.

Verification. The PR's check results and the changelog item numbers (2.1.328-124 and -125) are the author's claims, and I did not check them. I ran no commands and fetched neither linked page.

Spend. This review run used prepaid credits only.

Verifier dropped 0 of 0 candidates.

@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/address-feedback

Run 44951822055, attempt 1

Field Value
Model rung 0
Leg handoff
Run job success
Activity step success
Commits pushed none (read lane)
Run 44951822055

Failed GitHub API calls

No failed call recorded by the run job's scripts.

Lane-Activity: pr-unblock/address-feedback
Lane-Thread: PRRT_kwDOTCGFQM66bSE9
@github-actions

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/apply-feedback-patch

Run 44940316630, attempt 1

Field Value
Model rung 0
Leg base
Run job success
Activity step success
Commits pushed 87c556c
Run 44940316630

Failed GitHub API calls

No failed call recorded by the run job's scripts.

@melodic-automation-lanes

Copy link
Copy Markdown
Contributor Author

pr-unblock / address-feedback finished round 1 of max 4 on 87c556c.

@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/publish-feedback

Run 44951822055, attempt 1

Field Value
Model rung 0
Leg base
Run job success
Activity step success
Commits pushed none recorded
Run 44951822055

Failed GitHub API calls

No failed call recorded by the run job's scripts.

@melodic-automation-lanes

Copy link
Copy Markdown
Contributor Author

Unattended merge: native auto-merge leaves this pull request unarmed at head 87c556c7597adbc9e3309045d7286da274c6ec8f. No person needs to act: a lane clears this reason, and the next scheduled run judges it again.

pr-merge's gate refuses its verifier verdict (verify:fail)

…ation

pr-verify fail note PRRT_kwDOTCGFQM66bd79: #17908's first done-when item
asks the record to state the current /code-review scope and the Claude
Code version it was read from. The extraction record now states it as of
2.1.328, read 2026-11-21 from changelog items 2.1.328-124 and
2.1.328-125: on most models it reports only bugs and CLAUDE.md
violations, sends reuse, simplification and efficiency cleanups to
/simplify, and marks findings Important, Pre-existing or Nit. The live
commands-row pointer stays for anything newer.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RAoAtWJiWJdXQiPG8Xswsu

@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 be91a01d4a0e3f0170fab2446dce947feb58bbb1; the summary is in this lane's pull request comment.

claude-lane: claude-review

Comment on lines +42 to +45
That menu description is history, not the current scope. The scope as of Claude Code 2.1.328, read
2026-11-21 from that release's changelog (items 2.1.328-124 and 2.1.328-125): on most models
`/code-review` reports only bugs and CLAUDE.md violations, sends reuse, simplification and
efficiency cleanups to `/simplify`, and marks each finding Important, Pre-existing or Nit. Nothing

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.

This brings back the restated upstream scope that the previous push removed. The sentence from "The scope as of Claude Code 2.1.328, read 2026-11-21…" through "…Important, Pre-existing or Nit." paraphrases what the bundled /code-review reports. It also copies its severity labels.

.claude/rules/skill-bodies-state-current-rules.md applies to plugins/*/skills/**. It says:

A skill or agent body never restates a volatile upstream specific, quoted or paraphrased.

A record never cures a restatement: restated text with an as-of stamp or Claim/Basis block beside it is still the defect, fixed by deleting the restated value and keeping the pointer.

The new sentence breaks the rule whichever class it falls in:

  • As described behavior ("reports only bugs and CLAUDE.md violations", "sends … cleanups to /simplify"), the rule classes it as a pointer. A pointer carries no date, yet this one has "read 2026-11-21".
  • As copied strings (Important, Pre-existing, Nit), it is a stored value. A stored value needs a recheck trigger, and none is given.

The paragraph also contradicts itself. Its last sentence still says to read what the bundled skill "reports now" live from the commands row, because nothing in this lane acts on that scope. So the copy can only drift. The 2.1.263 text this PR is retiring is exactly that kind of drift.

scripts/check-as-of-recheck.sh won't catch this. Its stamp pattern needs as of YYYY-MM-DD, and here the version sits between "as of" and the date.

The owner asked for this sentence in a reply on another thread. This lane's earlier thread on the same text is marked lines changed since resolution: yes, so I'm raising it again. To settle it, do one of these:

  • (a) Drop the sentence and keep the pointer:
    That menu description is history, not the current scope. Nothing in this lane acts on the bundled
    skill's scope. When someone needs to know what it reports now, …
  • (b) Keep it as an owner decision. Record an exception for this file in docs/conventions/instruction-exception-register/ or in the rule, so the next review doesn't flag it again.

If you choose (b), update the PR body to match the head. Its "Fix" section says the paragraph has "no as-of date, no recheck trigger and no paraphrase of upstream behavior".

claude-lane: claude-review

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

pr-unblock / address-feedback: UNCERTAIN, left open

Evidence: plugins/review/skills/code-review/reference/bundled-code-review.md:42, .claude/rules/skill-bodies-state-current-rules.md:10, .claude/rules/skill-bodies-state-current-rules.md:30

The finding reads correctly against the rule. Lines 42-45 paraphrase the bundled /code-review scope as of 2.1.328 and copy its severity labels. The rule says a skill body never restates a volatile upstream specific, and that adding a date or record beside a restatement does not fix it. The paragraph's last sentence also still sends readers to the live commands row. I did not apply either fix because both need a person. The head commit (be91a01d, 'record the 2.1.328 /code-review scope as a dated observation') added this sentence, and the finding says the owner asked for it on another thread. This thread carries no direction from a trusted person, so dropping the sentence (option a) would undo an owner's request on a bot's say-so. Option (b) needs an exception entry in docs/conventions/instruction-exception-register/README.md, and check-paths refuses that file. Owner, please choose: (a) delete the sentence from 'The scope as of Claude Code 2.1.328' through 'Important, Pre-existing or Nit.' and keep the pointer, or (b) record an exception and make the PR body's Fix section match the head.

@melodic-automation-lanes

Copy link
Copy Markdown
Contributor Author

pr-unblock / address-feedback finished round 2 of max 4 on be91a01.

@melodic-automation-lanes melodic-automation-lanes Bot added the needs-human Human-in-the-loop required; autonomous sessions must not resolve items carrying this. label Nov 21, 2026
@melodic-automation-lanes

Copy link
Copy Markdown
Contributor Author

The pr-unblock lane stopped on this pull request and needs a person: an adjudicator found that a disputed review finding turns on what the change is meant to do, which a person decides.

Reason: intent

  • Reason: feedback-question-intent
  • Resume key: be91a01d4a0e3f0170fab2446dce947feb58bbb1 (the head the lane stopped on)

To resume, deal with the reason above, remove the needs-human label (and do-not-merge, if set), then dispatch the pr-unblock workflow with pr-number=18023 and activity=address-feedback:

gh workflow run pr-unblock.yml -R melodic-software/claude-code-plugins -f pr-number=18023 -f activity=address-feedback

@github-actions

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/report-escalation

Run 44951822055, attempt 1

Field Value
Model rung 0
Leg base
Run job success
Activity step success
Commits pushed none recorded
Run 44951822055

Failed GitHub API calls

No failed call recorded by the run job's scripts.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

needs-human Human-in-the-loop required; autonomous sessions must not resolve items carrying this.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant