Content
53%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 second half of the body is a solid, actionable tool reference with executable examples and a clear output schema, but it is undermined by ~60 lines of auto-generated filler that duplicates the description and pads with generic process rules, plus path inconsistencies and one fabricated expected-output block. Trimming the boilerplate and unifying script paths would materially improve it.
Suggestions
Delete the boilerplate header block (When to Use, Key Features, Dependencies, Example Usage run plan, Implementation Details) — it repeats the frontmatter description verbatim and states generic process rules Claude already knows, nearly halving the file while losing nothing.
Unify script paths: the usage examples invoke '.trae/skills/gene-info/scripts/fetch_gene_info.py' while the validation and example-plan sections use 'scripts/fetch_gene_info.py'; pick one relative path and use it everywhere.
Fix the 'Quick Validation' section: replace the fabricated 'Expected output format' block (result file / PASS-FAIL summary) with what 'python scripts/fetch_gene_info.py --help' actually prints, and add a concrete post-run checkpoint for batch runs (e.g. verify each gene in the input list appears in the output JSON).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The first ~60 lines are template boilerplate: the frontmatter description is repeated verbatim in 'Key Features', and sections like 'When to Use' ('when the request matches its documented task boundary') and circular pointers ('See ## Usage above for related details', 'See ## Prerequisites above') add no skill-specific information. The Dependencies and Example Usage sections also duplicate content, and generic sections like 'Deterministic Output Rules' and 'Completion Checklist' state what Claude already knows. | 2 / 5 |
Actionability | Provides concrete, mostly copy-paste-ready commands for single gene, keyword search, batch, file input, export, and advanced flags (--include-sequence, --include-homology), plus a documented output field list. Minor gaps: inconsistent script paths ('.trae/skills/gene-info/scripts/...' vs 'scripts/...') and a 'cd "20260316/scientific-skills/..."' example path that won't exist for users, and the 'Expected output format' block under Quick Validation does not match what '--help' actually prints. | 4 / 5 |
Workflow Clarity | The usage sections sequence single → keyword → batch → file → export clearly, and validation content exists (Quick Validation, validate_skill.py, Recommended Workflow step 4, Completion Checklist). However the checkpoints are generic boilerplate rather than concrete verification of the batch results, and the quick-validation 'Expected output format' (a gene_info_result.md with PASS/FAIL summary) is fabricated relative to the documented script behavior, so validation is present but confusing. | 3 / 5 |
Progressive Disclosure | Scored against the actual bundle (scripts/fetch_gene_info.py and scripts/validate_skill.py both exist and are referenced with clear paths; no references/ or assets/ dirs). Sections are well-organized with headers and the tool documentation is appropriately inline for its size. Minor gaps: the conflicting '.trae/skills/...' vs 'scripts/...' path prefixes hurt navigation, and the boilerplate front section buries the real content ('# Gene Information Tool' starts at line 60). | 4 / 5 |
Total | 13 / 20 Passed |