Content
92%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is an exceptionally lean, well-sequenced refactoring playbook with concrete tools and strong validation feedback loops. Its main weakness is progressive disclosure: it points to several resource files that are not present in the bundle.
Suggestions
Ship the referenced resource files (resources/definition.md, resources/measurement.md, resources/governance.md) or inline the essential thresholds and git-forensics one-liners they promise, so the skill is self-contained.
Resolve the ../_shared/core/context-loading.md and quality-principles.md references — either bundle them or drop the links if the shared context is unavailable in this package.
Consider adding one short worked example (e.g. an extract-method before/after with the verify-then-commit sequence) to ground the abstract workflow in a concrete trace.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, directive prose that assumes Claude's competence — it does not explain what a god class, characterization test, or Hyrum's law is; every line earns its place. | 3 / 3 |
Actionability | Names specific executable tools and commands — `uvx lizard`/`uvx radon`, Serena MCP `rename_symbol`, jscodeshift/OpenRewrite/ast-grep/comby codemods, and `refactor:` commit type — making the workflow copy-paste ready. | 3 / 3 |
Workflow Clarity | Clear sequenced Scenes (PREPARE/ACQUIRE/REASON/ACT/VERIFY/FINALIZE) with explicit validation checkpoints ('Re-run existing tests UNCHANGED -> commit') and a failure/recovery table including Mikado revert feedback loops for risky production changes. | 3 / 3 |
Progressive Disclosure | References to resources/definition.md, resources/measurement.md, resources/governance.md, and ../_shared/core/*.md are clearly signaled with one-line summaries, but none of those bundle files actually exist in the skill directory. | 2 / 3 |
Total | 11 / 12 Passed |