Content
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill attempts to cover a broad QA domain but suffers from verbosity and lack of concrete executable examples. The mutation testing section, while informative, dominates the document and includes explanatory content Claude doesn't need (what mutation testing is, why it matters). The core QA responsibilities are described at a checklist level without actionable code templates for writing actual tests.
Suggestions
Move the mutation testing tools table and detailed workflow to a separate reference file (e.g., docs/mutation-testing.md) and keep only a 3-line summary with a link in the main SKILL.md
Add concrete, executable test examples for each test type (unit, integration, E2E) using the mentioned frameworks (Vitest/Jest/Playwright) — even one example per type would dramatically improve actionability
Remove explanatory content Claude already knows (what mutation testing is, why coverage matters, the Anti-Rationalization table) to cut token usage by ~40%
Add an explicit unified workflow with validation checkpoints: e.g., '1. Run existing tests → 2. If failing, fix first → 3. Write new tests for uncovered ACs → 4. Verify coverage threshold → 5. If below, add tests → 6. Generate evidence report'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is excessively verbose at ~200+ lines. It explains concepts Claude already knows (what mutation testing is, what coverage means, what PDF-like analogies are), includes lengthy tables of tools per language that could be a reference file, and has significant padding like the 'Anti-Rationalization' table and philosophical explanations about why tests matter. Much of this content doesn't earn its token cost. | 1 / 3 |
Actionability | There is one concrete code example (SQLite WAL cleanup pattern) and some specific commands (e.g., `npx stryker run`), but most guidance is abstract checklists and general principles rather than executable code. The mutation testing workflow uses numbered prose steps rather than concrete commands, and the core QA sections (unit, integration, E2E) lack any executable test examples or templates. | 2 / 3 |
Workflow Clarity | The mutation testing workflow has a reasonable sequence but lacks explicit validation checkpoints and error recovery loops. The main QA workflow is scattered across multiple sections (checklist, coverage, handoff) without a clear unified sequence. There's no explicit 'if tests fail, do X' feedback loop for the core testing workflow, and the checklist is more of a gate than a process. | 2 / 3 |
Progressive Disclosure | References to external files exist (personas/test-engineer.md, docs/skill-guides/qa-testing.md, various policies), which is good. However, the main SKILL.md itself is monolithic with too much inline content that should be in reference files (the entire mutation testing tools table, the anti-rationalization table). The mutation testing section alone is nearly half the document and could be a separate reference. No bundle files were provided to verify reference accuracy. | 2 / 3 |
Total | 7 / 12 Passed |