fix(review,init): single severity vocab, BASE_REF capture, command -v parallel-cancel

[slbug]

No description provided.

[Copilot]

Pull request overview

This PR updates the Ruby/Grape/Rails Claude Code plugin’s review/init instructions and release metadata for v1.16.13, focusing on review severity vocabulary consistency, safer base-ref capture guidance, and tool detection behavior.

Changes:

• Standardizes review severity wording around blocker | warning | suggestion for workers and BLOCKER | WARNING | SUGGESTION for consolidated output.
• Moves /rb:review base-ref/diff collection guidance into the skill body and adds artifact-reading guidance.
• Bumps release metadata to 1.16.13 with a matching changelog entry.

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
plugins/ruby-grape-rails/skills/review/SKILL.md	Updates review collection, fanout, severity, and artifact-reading instructions.
plugins/ruby-grape-rails/skills/review/references/review-playbook.md	Aligns worker severity mapping and synthesis guidance with the new vocabulary.
plugins/ruby-grape-rails/skills/init/SKILL.md	Replaces multi-arg command -v guidance with independent tool probing.
plugins/ruby-grape-rails/references/preferences/tool-batching.md	Adds BAD/GOOD guidance for multi-tool command -v detection.
plugins/ruby-grape-rails/agents/migration-safety-reviewer.md	Updates sample severity categories to blocker/warning/suggestion.
plugins/ruby-grape-rails/agents/iron-law-judge.md	Updates violation/severity examples and confidence casing.
plugins/ruby-grape-rails/agents/data-integrity-reviewer.md	Updates sample severity categories to blocker/warning/suggestion.
plugins/ruby-grape-rails/.claude-plugin/plugin.json	Bumps plugin version to 1.16.13.
package.json	Bumps package version to 1.16.13.
package-lock.json	Bumps lockfile package versions to 1.16.13.
CHANGELOG.md	Adds the 1.16.13 release notes and version links.
.claude-plugin/marketplace.json	Bumps marketplace metadata and plugin ref to v1.16.13.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 17 changed files in this pull request and generated 9 comments.

Comments suppressed due to low confidence (1)

plugins/ruby-grape-rails/agents/verification-runner.md:105

• This second Pronto instruction has the same shell-scope problem: $BASE_REF only exists in the shell that evaluated resolve-base-ref, so a separate Bash call for bundle exec pronto will run with an empty/unset base ref. Keep the eval and Pronto command in one Bash invocation or state that explicitly.

- If `PRONTO_AVAILABLE=true`, run
  `eval "$(${CLAUDE_PLUGIN_ROOT}/bin/resolve-base-ref)"` (populates
  `$BASE_REF`, `$REMOTE`, `$DEFAULT_BRANCH`); then
  `bundle exec pronto run -c "$(git merge-base HEAD "$BASE_REF")"`
  as the optional final diff-scoped step

[Copilot]

Pull request overview

Copilot reviewed 17 out of 19 changed files in this pull request and generated 7 comments.

[Copilot]

Pull request overview

Copilot reviewed 48 out of 50 changed files in this pull request and generated 12 comments.

Comments suppressed due to low confidence (6)

lab/eval/output_checks.py:849

• This second At-a-Glance parser also normalizes with .title(), so non-canonical severity casing is accepted during the coverage cross-check. Keep this strict so the validator rejects old BLOCKER/WARNING/SUGGESTION artifacts under the new title-case contract.

        severity = row[2].strip().title()
        reviewer = row[4].strip()
        new_state = row[6].strip().lower()
        if severity not in {"Blocker", "Warning", "Suggestion"}:
            bad_enum.append(
                f"At-a-Glance row {idx}: `Severity` cell {row[2]!r} not in "
                "{`Blocker`, `Warning`, `Suggestion`} — malformed enum"

plugins/ruby-grape-rails/skills/triage/SKILL.md:260

• The phase-output instructions still use the old uppercase NEW WARNINGs / NEW SUGGESTIONs vocabulary. This conflicts with the new title-case plan/template examples and can train triage to emit mixed casing in generated plans.
  plugins/ruby-grape-rails/skills/triage/SKILL.md:294
• This paragraph still uses BLOCKERs even though the surrounding shortcut labels were converted to title case. Use the same Blocker/Blockers vocabulary here so the selection UI instructions do not mix old and new bucket names.
  plugins/ruby-grape-rails/skills/review/SKILL.md:354
• This repeats the new bulk-cat/Read rule a second time in the same shipped skill body. Keep the detailed tool-batching guidance in the preference reference and point there instead of duplicating it in the review workflow text.
  plugins/ruby-grape-rails/skills/triage/SKILL.md:360
• The all-blocker edge-case sample still emits BLOCKERs while the condition now uses Blocker. This canonical UI text should use the same title-case vocabulary to avoid training /rb:triage to produce mixed casing.
  plugins/ruby-grape-rails/skills/full/references/example-run.md:69
• Only this line was converted to Suggestion; the same example block still uses old uppercase SUGGESTIONs, BLOCKERs, and WARNINGs nearby. This keeps the shipped /rb:full example split across two severity vocabularies despite the new review casing contract.

[Copilot]

Pull request overview

Copilot reviewed 48 out of 50 changed files in this pull request and generated 7 comments.

Comments suppressed due to low confidence (3)

plugins/ruby-grape-rails/skills/review/SKILL.md:355

• This Gotcha restates the Read-over-cat/tool-batching discipline inside a shipped skill body. The repo policy requires skill bodies to use pointer-only links to the preference doc rather than duplicating the rule text and rationale.
  plugins/ruby-grape-rails/skills/triage/SKILL.md:357
• The heading and condition use singular "Blocker" for a plural set of findings; use the plural form here to match the following "All 5 findings are Blockers" sentence.
  plugins/ruby-grape-rails/skills/triage/SKILL.md:298
• The shortcut names use singular nouns with "Select All" even though the operation selects multiple findings; plural labels would read correctly and match the surrounding plural "Blockers" wording.

[Copilot]

Pull request overview

Copilot reviewed 48 out of 50 changed files in this pull request and generated 13 comments.

Comments suppressed due to low confidence (2)

plugins/ruby-grape-rails/agents/iron-law-judge.md:101

• The section says all Iron Law violations are Blockers, but the Blocker table below still omits Laws 5, 8, 9, and 17 from the 22-law registry above. Those omitted laws can be treated as lower-priority or missed entirely by this agent despite the new all-Blocker doctrine.

| Law | Pattern | Risk |
|-----|---------|------|
| 1 | `t.float :price` | Financial errors |
| 2, 15 | `where("id = #{id}")` | SQL injection |
| 3 | N+1 queries (loop without `includes`) | Performance / DB load |

plugins/ruby-grape-rails/agents/iron-law-judge.md:158

• This priority summary excludes Law 17 even though the section above says all Iron Law violations are Blockers. Law 17 remains in the canonical registry, so omitting it here creates an ambiguous priority for supervised-background-process violations.

1. **Blockers** (Laws 1-16, 18-20): All Iron Law violations block merge
2. **Verification** (Law 21): Testing discipline — Required
3. **Surgical Changes** (Law 22): Out-of-scope edits flagged separately

[Copilot]

Pull request overview

Copilot reviewed 50 out of 52 changed files in this pull request and generated 12 comments.

Comments suppressed due to low confidence (2)

plugins/ruby-grape-rails/skills/review/references/review-playbook.md:651

• This command is shown as a runnable diff-LOC calculation, but <MERGE_BASE> is shell redirection syntax when copied into Bash. Use an executable form or move the placeholder substitution instruction outside the command block so agents do not run an invalid command.
  lab/eval/output_checks.py:795
• The strict title-case severity enum introduced here is not covered by a regression test that rejects the old uppercase values (BLOCKER, WARNING, SUGGESTION). Add a negative test so the intended “no backward compat” behavior cannot be accidentally loosened by reintroducing normalization.

        severity = row[2].strip()
        new_state = row[6].strip().lower()
        if severity not in new_counts:
            bad_enum.append(

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 69 out of 72 changed files in this pull request and generated no new comments.

[slbug]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. 🚀

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".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".