Content
71%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 well-organized, lean instruction body with a clear workflow and concrete guidance, undermined by a missing rules/ bundle: all nine referenced rule files do not exist, breaking the progressive-disclosure structure it depends on.
Suggestions
Ship the referenced rules/ bundle (finding-features.md, naming.md, assertions.md, endpoint-tests.md, test-data.md, isolation.md, security.md, performance.md, review.md) so the Rule Index links resolve.
Add a brief fix-and-retry loop to 'How to Apply' (e.g., if tests fail, fix the code or test and re-run the minimal set until green) to strengthen the validation checkpoint.
Tighten the Consistency First section, which restates the convention-precedence rule three times, to recover a few tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient and free of basic-concept padding, but the Consistency First section restates the convention-precedence idea three ways and a few sentences could be tightened. | 4 / 5 |
Actionability | Concrete, specific guidance ('Write a feature test first. Write a unit test only for logic that does not use the framework.'; 'Run the smallest set of tests that covers the change.') with only minor gaps — e.g., no example test command or assertion snippet inline. | 4 / 5 |
Workflow Clarity | The 5-step 'How to Apply' sequence has explicit checkpoints (tests must pass; resolve every mismatch before completion) and a final checklist, but lacks a structured fix-and-retry loop when tests fail. | 4 / 5 |
Progressive Disclosure | The SKILL.md is a lean overview with a well-signaled one-level-deep Rule Index table — good structure — but every referenced rules/*.md file is absent from the bundle, so navigation leads nowhere. | 3 / 5 |
Total | 15 / 20 Passed |