Content
82%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 tight, highly actionable skill body: three complete test templates, a clear decision structure by module type, and an exact run command, with essentially no wasted tokens. The only notable gaps are a missing `vi` import in the component template and the absence of an explicit fix-and-re-run loop after running vitest.
Suggestions
Fix the component example's imports: it uses `vi.fn()` but the import line is `import { describe, it, expect } from 'vitest'` — add `vi` to make the template copy-paste ready.
Add a brief step 5 for the failure path, e.g., 'If tests fail, fix the test or mock and re-run' — this closes the feedback loop and would lift workflow clarity to a 5.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean with zero explanation of concepts Claude already knows — it opens directly with "Create tests for the specified file or module" and every subsequent token is a template, a file-path convention, or a command. This matches the anchor-5 'lean and efficient; every token earns its place' example. | 5 / 5 |
Actionability | Provides complete code templates for all three test types, an exact output path convention (`__tests__/OriginalName.spec.ts(x)`), and a specific run command (`npx vitest run --reporter=verbose`). It misses anchor 5's copy-paste-ready bar because the component example uses `vi.fn()` without importing `vi` from vitest, and the templates rely on xxx placeholders without guidance on adapting them. | 4 / 5 |
Workflow Clarity | The four steps (analyze → determine test type → create file → run vitest) form a clear sequence, and step 4's test run serves as a checkpoint — matching anchor 4's 'clear sequence with most checkpoints present'. There is no explicit feedback loop for handling failures (fix and re-run), which anchor 5 requires; the destructive/batch cap does not apply since test creation is non-destructive. | 4 / 5 |
Progressive Disclosure | The body is well-organized with a `## Steps` structure and clearly labeled per-type subsections, and no external references are needed. At ~85 lines it exceeds the under-50-line simple-skill exception, and the three full code templates are inlined, so it sits at anchor 4 ('good structure; most content appropriately placed') rather than 5. | 4 / 5 |
Total | 17 / 20 Passed |