Content
90%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 tight, highly actionable reference skill: executable verified commands, gnomAD-specific data semantics only where needed, and clean sections. The only gaps are the absence of query-failure/validation guidance and an unreferenced bundled citation file.
Suggestions
Add a brief validation step after each query, e.g. 'Check the output JSON for an "error" key or empty results before reporting frequencies; retry after the script's rate-limit backoff if the request failed.'
Reference the bundled citation in the References section (e.g. 'Cite gnomAD as described in references/citation.bib when using results in reports') so the file is discoverable from SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and operational: prerequisites, two core rules, three command templates, and two reference URLs. The only explanatory material (the FAF definition and the 'oe_lof_upper' field mapping for LOEUF) is gnomAD-specific API/data semantics Claude would not know, so every token earns its place. It does not match the level-4 anchor since there is no over-explanation to trim. | 5 / 5 |
Actionability | Fully executable, copy-paste-ready 'uv run' commands for all three scripts with real flags, output paths, and input-format examples ('1-55516888-G-GA', 'rs1800562'); the documented flags match the actual script CLIs in the bundle. Matches the top anchor of executable commands covering the common cases. | 5 / 5 |
Workflow Clarity | Prerequisites are sequenced with a conditional checkpoint (license file exists → notify → record), and each query usage is unambiguous. However, there is no validation or error-recovery guidance for query outcomes (e.g., inspecting the output JSON for API errors or empty results), so it fits 'clear sequence with most checkpoints present; minor validation gaps' rather than the top anchor requiring explicit validation steps and feedback loops. | 4 / 5 |
Progressive Disclosure | Good structure: clear sections, scripts clearly signaled with absolute paths and per-script usage, and external documentation links. The body is ~60 lines (outside the under-50-line exception), and the bundled references/citation.bib is never mentioned in SKILL.md, so one reference is undiscoverable — a minor organization gap that fits the level-4 anchor rather than 'easy navigation' with all references well-signaled. | 4 / 5 |
Total | 18 / 20 Passed |