Content
92%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 exemplary lean skill body: a crisp numbered workflow with genuine validation checkpoints, a failure feedback loop, and a done-checklist, plus a clearly signaled reference that exists in the bundle. The only minor weakness is the generic "Run the project's test command" phrasing.
Suggestions
Make the test-run step more concrete, e.g. "Run the project's test command (e.g. npm test / pytest) — discover it if unknown" to reduce the one generic instruction.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | "Write the test first. Watch it fail. Write the minimal code to pass. Refactor with the tests green." — every line is directive with zero re-teaching of concepts Claude already knows. Not 4: there is no over-explanation anywhere to trim; even the rationale lines ("If it passes, it tests nothing new") are actionable rules, not padding. | 5 / 5 |
Actionability | The cycle gives concrete, specific guidance ("Write one test for one behavior", "The test must fail because the behavior is missing, not because of a typo or a missing import"), and test-writing detail is correctly delegated to the reference. Not 5: "Run the project's test command" is generic rather than a specific command or example, leaving minor gaps in executable specificity. | 4 / 5 |
Workflow Clarity | The Red/Green/Refactor cycle is a clearly numbered sequence with explicit validation checkpoints ("Watch it fail", "Watch it pass", "Run the full suite"), a failure feedback loop ("If it passes, it tests nothing new: change the test"), and a final "Before you say it is done" checklist. This matches the top anchor including error-recovery feedback. | 5 / 5 |
Progressive Disclosure | A short, well-sectioned body with a single one-level-deep reference — "When writing or changing a test, read [references/writing-good-tests.md]" — which points to a real, relevant file in the bundle. Content is appropriately split (process inline, test-writing craft in the reference); nothing is inlined that belongs elsewhere. | 5 / 5 |
Total | 19 / 20 Passed |