Content
85%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.
A rigorous, well-structured review skill with explicit gates, a sequenced workflow, and clean one-level-deep references. Its only real weakness is raw length, which pressures conciseness slightly.
Suggestions
Tighten the Core Contracts and Scope and Evidence sections — several rule sentences restate the same review-only constraint and could be merged to cut tokens without losing meaning.
Consolidate the repeated 'Completion Gate / Review Incomplete' gating logic that appears in both Review Incomplete Proof Gate and Completion Gate and Scripts to avoid restating the same terminal-classification rule.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Information-dense and assumes Claude's competence — no padding with what a PR/CI is — but the body runs long and a few rule sections could be tightened without losing clarity. | 4 / 5 |
Actionability | Provides concrete executable commands ('scripts/review_ledger.py validate-incomplete --ledger <ledger>', 'scripts/build_review_inventory.py --format json'), specific file paths, and an explicit decision/output contract, with only minor gaps. | 4 / 5 |
Workflow Clarity | The Review Workflow is a clearly sequenced 1–7 process with explicit validation gates (Finding Proof Gate, Review Incomplete Proof Gate, Completion Gate) and a convergence feedback loop ('return to step 5 and repeat'). | 5 / 5 |
Progressive Disclosure | SKILL.md is an overview with well-signaled one-level-deep references to evidence-access.md, high-risk-review.md, sql-parser-review.md, and review-corrections.md, all verified present; detailed material is appropriately split out. | 5 / 5 |
Total | 18 / 20 Passed |