Content
48%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.
The body contains a genuinely sound refactoring safety workflow — green baseline, incremental atomic changes, test after each step, commit checkpoints, and a revert-on-failure recovery protocol — but buries it under textbook knowledge catalogs (SOLID, design patterns, code smells) Claude already knows and generic template filler sections. The code example is illustrative rather than executable, and no operational commands (e.g., the actual test command) are specified.
Suggestions
Cut the textbook catalogs (SOLID definitions, GoF pattern lists, code-smell taxonomy) — Claude already knows these; keep only guidance on which techniques to prefer and in what order, or move them to a references/ file with clear pointers.
Remove the generic template filler (the circular "Use this skill when" / "Do not use this skill when" bullets and the boilerplate Instructions bullets) and fix the unclosed Prompt quote in the Usage section.
Specify the actual verification commands (e.g., the project's test runner invocation) instead of the generic "Run tests", and either make the TypeScript example self-contained or trim it to the essential diff.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Quotes: "Single Responsibility: One reason to change", "Apply Creational (Factory, Builder, Singleton)", "Working on tdd workflows tdd refactor tasks or workflows" — multiple sections catalog textbook concepts Claude already knows (SOLID, GoF patterns, code smells) alongside circular template filler. Terse bullet formatting keeps it above a 1, but the padding is far more than "some". | 2 / 5 |
Actionability | Quotes: "Duplicated code → Extract methods/classes", "Run tests to establish green baseline", "Profile to identify bottlenecks" — the smell→technique mappings and the full before/after example are concrete, but no executable commands are given (which test command? which profiler?) and the example references undefined types (OrderTotal, ValidationResult) with an incomplete Before snippet, leaving key details missing. | 3 / 5 |
Workflow Clarity | Quotes: "Run tests to establish green baseline", "Make small, atomic changes", "Run tests after each modification", "If tests fail: Immediately revert last change" — a clear sequence with explicit validation checkpoints, a safety checklist, and a revert-on-failure recovery loop. It falls short of a 5 because the process is diluted across ten parallel catalog sections and the unclosed Prompt quote in Usage muddies the entry point. | 4 / 5 |
Progressive Disclosure | The body runs ~185 lines with section headers (some structure), but the design pattern, SOLID, and advanced-pattern catalogs are inlined content that clearly belongs in a separate reference file, and there are no external file references at all — matching the anchor of content that should be separate sitting inline. | 3 / 5 |
Total | 12 / 20 Passed |