Content
88%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An exceptionally actionable, well-sequenced operational skill: every step is executable, guarded by explicit validation and stop conditions, and closed out with a required output format. The only refinements are minor trimming of explanatory asides and potentially splitting a reference table out of the main file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational — commands, tables, and rules with essentially no explanation of concepts Claude already knows. Minor instances of over-explanation could be trimmed, e.g., the rationale for treating local and third-party reviewers as complementary in step 2 and some long parentheticals in hard rule 4. | 4 / 5 |
Actionability | Fully executable throughout: concrete `gh pr view`/`gh api`/`git` commands, copy-paste-ready commit-message heredoc, exact GraphQL mutation for resolving threads, and a triage table mapping each failure classification to a specific Conventional Commits type. Placeholder variables ($PR, $PR_BRANCH, $COMMENT_ID) are all defined by prior commands. | 5 / 5 |
Workflow Clarity | A clear six-step sequence with explicit validation checkpoints and feedback loops: prerequisite guards that abort with exit codes before any action, per-failure diagnostic commands (`gh run view --log-failed`), hook-blocking push guidance ('fix the cause, don't reach for the bypass'), bounded polling, and five explicit stop conditions with escalation paths. This is exactly the validate → fix → retry structure the anchor describes. | 5 / 5 |
Progressive Disclosure | No bundle files exist, and the skill is appropriately self-contained: well-organized sections with a clear procedure, output format, and a Cross-references section pointing one level deep to sibling skills (/review-pr, /implement, /diagnose, /merge-train) and repo records (constitution/local-workflow.md, docs/adr/INDEX.md). At ~230 lines some inline material (e.g., the failure-classification table) could arguably live in a reference file, which keeps it at 'good structure, minor gaps' rather than optimal split. | 4 / 5 |
Total | 18 / 20 Passed |