Content
72%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 lean, well-structured skill body that appropriately keeps its bulk in a single verified reference file. Its weaknesses are the absence of any validation/verification checkpoint in the Check-to-Fix flow and slightly abstract directives ('right level of abstraction') where concrete criteria or an inline snippet would help.
Suggestions
Add an explicit verification step to the workflow, e.g., after Fix: 'Re-run the affected test suite and confirm it fails when the mock is removed' — a feedback loop that catches over-mocked or broken tests.
Replace abstract directives like 'mocking at the right level of abstraction' with concrete criteria (e.g., 'mock at the network/module boundary, not inside business logic') or one inline example before deferring to the reference.
Trim the opening sentence that explains why over-mocking is bad — Claude already knows this — and keep only the rule framing needed for the review task.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is tight: terse section directives, a lean Quick Reference of actionable bullets, and no padded library comparisons or explanations of what jest/MSW are. The single framing sentence ('Proper mocking isolates units under test... false confidence') explains a concept Claude already knows and could be trimmed, which is exactly the 'minor instances of over-explanation' of score 4 rather than the every-token-earns-its-place of score 5. | 4 / 5 |
Actionability | Concrete, executable specifics are present: 'jest.restoreAllMocks() in afterEach', 'Use MSW', and named mock targets (APIs, databases, file system), plus specific check directives ('over-mocking, missing mock cleanup, proper mock implementation'). It falls short of score 5 because directives like 'mocking at the right level of abstraction' and 'proper mock implementation' are abstract, and no inline example is given — code examples live entirely in the reference file. | 4 / 5 |
Workflow Clarity | The sections (Quick Reference, Check, Fix, Explain, Code Review) read as parallel modes with only an implicit Check-then-Fix sequence, and there are no validation or verification checkpoints anywhere (e.g., re-run tests after cleanup changes, confirm the CI check actually fails on a regression). This is not a destructive/batch operation so the hard cap does not apply, but the missing checkpoints match anchor 3; it is not score 4 because no explicit sequence or feedback loop is present. | 3 / 5 |
Progressive Disclosure | The ~35-line body is a clean overview: each section is short, and full implementation details are clearly signaled one level deep via 'For full implementation details, code examples, and framework-specific guidance, see references/rule.md', which exists as a real 165-line file containing exactly those code examples. This matches the anchor of a clear overview with a well-signaled, one-level-deep reference and easy navigation. | 5 / 5 |
Total | 16 / 20 Passed |