Content
86%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.
An exemplary compact instruction skill: every line is directive, the restart-bugs heuristic is genuinely non-obvious knowledge, and the structure suits its size. The only refinements are trimming the rationale text and making fix verification an explicit final step. No bundle files exist, and none are needed at this scope.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and punchy ("Code doesn't change between runs. State does."), assuming Claude's competence throughout. Not 5: the "Why:" paragraph and bullet parentheticals restate rationale Claude largely already knows and could be trimmed; not 3: nothing is noticeably verbose or padded. | 4 / 5 |
Actionability | Concrete, executable instruction-only guidance: "grep for the same pattern, fix all instances", "add logging, read the actual error", and named artifacts for restart bugs (config files, caches, lock files, serialized state). Per the scoring notes, absence of code is not penalized when guidance is this specific and directive. | 5 / 5 |
Workflow Clarity | The Pattern is a well-ordered checklist (reproduce → ask why → resist guards → grep pattern → instrument) with a validation seed ("if you can't reproduce it, you can't verify your fix"). Not 5: verifying the fix is implicit rather than an explicit checkpoint step; not 3: sequence and most checkpoints are present. | 4 / 5 |
Progressive Disclosure | A sub-50-line, single-purpose skill with no need for external references and well-organized sections (title, Why, Pattern, restart-bugs subsection). Per the simple-skill guideline, this scores 5 with just well-organized sections. | 5 / 5 |
Total | 18 / 20 Passed |