Content
67%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, opinionated methodology skill with strong actionable directives (seam agreement, loop rules, explicit question scripts) and clean structure. Its main costs are re-explaining well-known testing anti-patterns and pointing to example files that are absent from the bundle.
Suggestions
Trim the Anti-patterns section to the rule plus the tell (e.g., 'Implementation-coupled — the test breaks when you refactor but behavior hasn't changed') without re-teaching what each anti-pattern is or worked examples like expect(add(a, b)).toBe(a + b).
Ship tests.md and mocking.md in the bundle or inline the few needed examples — both are referenced for the skill's core illustrative content but do not exist in the evaluated bundle.
Make the red-step validation explicit: add a checkpoint such as 'Run the test and confirm it fails for the expected reason before writing any implementation.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The seams protocol ("Test only at pre-agreed seams", "No test is written at an unconfirmed seam") and the loop rules are genuinely additive, but the longest section — "Anti-patterns" — re-teaches concepts Claude already knows (implementation-coupled tests, tautological assertions with `expect(add(a, b)).toBe(a + b)`, horizontal slicing), and "What a good test is" restates the basic public-interface principle. Mostly efficient with a noticeable portion of known material, which fits the 3 anchor better than 4's "minor instances of over-explanation". | 3 / 5 |
Actionability | Concrete, executable directives throughout: "Before writing any test, write down the seams under test and confirm them with the user", a literal script to ask ("What's the public interface, and which seams should we test?"), "Write the failing test first, then only enough code to pass it", and an explicit tool call ("call the Skill tool with 'codebase-design'"). Only minor gaps: no inline example of a seam note or a good test, since examples are delegated to tests.md. | 4 / 5 |
Workflow Clarity | The red → green sequence is clearly stated with an explicit pre-loop checkpoint (seam agreement: "No test is written at an unconfirmed seam") and per-cycle discipline ("One seam, one test, one minimal implementation per cycle"). Not a 5 because the red-side validation is implicit — there is no explicit step to run the test and confirm it fails before implementing, only "Red before green." No destructive/batch cap applies. | 4 / 5 |
Progressive Disclosure | Short, well-sectioned overview with clearly signaled one-level-deep references ("See [tests.md](tests.md) for examples and [mocking.md](mocking.md) for mocking guidelines") plus well-scoped cross-references to the codebase-design and code-review skills. Not a 5 because the referenced tests.md and mocking.md do not exist anywhere in the bundle, so the signaled navigation breaks and the promised examples are unreachable. | 4 / 5 |
Total | 15 / 20 Passed |