Content
88%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 command-reference skill body: every section is executable guidance with concrete examples, and the workflows are sequenced with 'facts check' as an explicit validation loop. The only real gaps are minor alias redundancy and a single-file layout that inlines a few sections (domain vocabulary, writing guidance) that could live in one-level-deep reference files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is command-first and dense with no explanations of concepts Claude already knows, but carries minor trimmable redundancy — aliases are defined twice (the 'Recommended short aliases' block and again as inline 'or the alias' comments), plus editorial asides like 'These are the highest-ROI shortcuts'. | 4 / 5 |
Actionability | Fully executable throughout: copy-paste-ready commands with concrete arguments (e.g. the 'facts add' example with a real curl validation command), explicit alias definitions, a complete file-format spec, and a runnable boolean tag-filter expression. | 5 / 5 |
Workflow Clarity | Agent workflows are clearly sequenced (orient with 'facts ll'/'facts check' → define spec → track lifecycle → maintain accuracy) with 'facts check' serving as an explicit validation feedback loop at each stage, including guidance to verify facts before removing code and to lint-then-run during check. | 5 / 5 |
Progressive Disclosure | No bundle files exist and the single-file structure is well organized with clear sections, but at ~194 lines the Domain vocabulary, Writing good facts, and Companion skills sections are separable content that could move to reference files, and no reference pointers are used. | 4 / 5 |
Total | 18 / 20 Passed |