Content
82%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.
A well-structured, highly actionable skill body: executable commands for all common cases, explicit anti-patterns with validation and recovery guidance, and proper deferral of setup detail to a reference file. The main issues are a dead reference to a nonexistent requirements.txt, validation checkpoints separated from the workflow steps rather than integrated into them, and overlap between the Mindset and Anti-Patterns sections.
Suggestions
Fix the broken reference: 'Dependencies: [requirements.txt](requirements.txt)' points to a file that does not exist in the bundle — either add requirements.txt or remove the link and fold dependency listing into references/setup-and-troubleshooting.md.
Integrate validation checkpoints into the numbered Workflow steps (e.g., after step 3 add 'spot-check rows with --print before saving'; after step 7 add 'read the batch summary and re-run failed files'), so the checks are encountered in sequence rather than only in the Anti-Patterns section.
Deduplicate the Mindset and Anti-Patterns sections — scanned-PDF handling, batch non-atomicity, and output validation are each covered twice; keep one section for principles and point to the Anti-Patterns for the BAD/GOOD pairs.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient: copy-paste commands with inline comments, no explanations of concepts Claude already knows, and details deferred to a reference file. It fits the 4 anchor ('Efficient; minor instances of over-explanation that could be trimmed') — the Mindset section substantially restates the Anti-Patterns (scanned-PDF handling, batch non-atomicity, output validation each appear twice), which is more than the lean 5 anchor's 'every token earns its place' but well short of the 3 anchor's 'unnecessary explanation'. | 4 / 5 |
Actionability | Every workflow step is a fully executable command with real flags ('./scripts/extractor.py paper.pdf --template enzyme -o results.md', '--format csv', '--ocr mathpix', './scripts/batch_extract.py ./literature/ ./output/ --template enzyme --format csv'), and the GOOD custom-prompt example is copy-paste ready with exact column format. This matches the 5 anchor ('Fully executable; copy-paste ready code or commands; specific examples cover the common cases') — preset templates, custom prompts, CSV, Mathpix OCR, batch, and terminal printing are all covered. | 5 / 5 |
Workflow Clarity | The eight numbered steps form a clear install→configure→run sequence, and validation with feedback loops is explicitly present (spot-check 3 rows with --print, re-run failed batch files individually, switch to --ocr mathpix on empty extraction, check Mathpix credentials before invoking), so the batch-operation cap of 3 does not apply. It sits at the 4 anchor rather than 5 because those checkpoints live in the separate Anti-Patterns section instead of being sequenced into the workflow steps — a reader following only the Workflow section would miss the validation checkpoints, which is the 'minor validation gaps' of anchor 4. | 4 / 5 |
Progressive Disclosure | The SKILL.md is a lean overview with well-signaled, one-level-deep references ('See [setup-and-troubleshooting.md](references/setup-and-troubleshooting.md)', a References section listing Setup/Script/Upstream) and scripts appropriately separated, matching the 4 anchor ('Good structure; most content is appropriately placed; references mostly clear; minor organization gaps'). It falls short of 5 because the References section lists 'Dependencies: [requirements.txt](requirements.txt)' which does not exist in the bundle — a broken navigation pointer. | 4 / 5 |
Total | 17 / 20 Passed |