Content
56%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 skill delivers strong actionability with executable commands, a parameter reference, and explicit error-handling and validation guidance, and its reference bundle is well-signaled and one level deep. However, it is held back by heavy inlining of content duplicated in its own reference files and by generic boilerplate sections that inflate token cost without adding skill-specific guidance.
Suggestions
Replace the inlined NIH Scoring System tables and Common Weaknesses Detected lists with brief pointers to references/nih_scoring_rubric.md and references/common_weaknesses_catalog.md, keeping only a short summary of the 1-9 scale and the five criteria.
Trim the generic template sections (Risk Assessment, Security Checklist, Evaluation Criteria, Lifecycle Status, Output Requirements/Contract boilerplate) down to only the items that carry skill-specific guidance.
Merge the Quick Check and Audit-Ready Commands sections into one validation block, remove the duplicated '--help' invocation, and either include the referenced requirements.txt or drop the pip install prerequisite.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is noticeably verbose: the NIH scoring tables and the ~60-line Common Weaknesses catalog duplicate content already shipped in references/nih_scoring_rubric.md and references/common_weaknesses_catalog.md, and generic boilerplate sections (Risk Assessment, Security Checklist, Evaluation Criteria, Lifecycle Status, Output Contract) pad the file without skill-specific value. | 2 / 5 |
Actionability | Concrete, copy-paste-ready CLI commands cover the common cases (full review, section-only, focus, scores-only, compare) alongside a library usage example and a full parameter table. Minor gaps keep it below 5: the Prerequisites section references a requirements.txt that is not present in the bundle, and the Audit-Ready Commands block repeats 'python scripts/main.py --help' twice and duplicates the Quick Check. | 4 / 5 |
Workflow Clarity | The 5-step Workflow is clearly sequenced with a concrete validation entry point (py_compile quick check) and explicit Error Handling / Failure Handling sections describing fallback paths. Not 5 because the workflow steps themselves are somewhat abstract ('Validate that the request matches the documented scope') and there is no validate-fix-retry feedback loop. | 4 / 5 |
Progressive Disclosure | The References section clearly signals all five one-level-deep reference files, but substantial content that belongs in those files (the NIH scoring tables and the Common Weaknesses lists) is inlined in SKILL.md instead of being pointed to, matching the 'content that should be separate is inline' anchor. | 3 / 5 |
Total | 13 / 20 Passed |