Content
57%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 body provides concrete, executable commands for all three subcommands and is well-structured for a small single-script skill. However, it is padded with generic template sections (Key Features, Implementation Details, boilerplate When to Use bullets) that add tokens without adding skill-specific knowledge, and its run plan references a CONFIG block the script does not have.
Suggestions
Cut the template boilerplate sections (Key Features, Implementation Details, and generic When to Use bullets) that restate the description or give generic process advice.
Add one fully worked example, e.g., `python scripts/hgnc.py fetch "BRAF"` with the key fields of the expected output, instead of only placeholder commands.
Fix the run plan: remove the step about editing an in-file CONFIG block, which does not exist in scripts/hgnc.py, and make the verification step concrete (e.g., check the returned JSON for the expected hgnc_id/symbol).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Several sections are template boilerplate rather than skill-specific instruction: "Scope-focused workflow aligned to: …", "Use this skill when a evidence insight task needs a packaged method instead of ad-hoc freeform output", and the generic "Implementation Details" bullets ("validate the request, choose the packaged workflow, and produce a bounded deliverable"). It is not 3 because the padding spans multiple whole sections (Key Features, Implementation Details, parts of When to Use and Dependencies), not just isolated spots; not 1 because the Tools section itself is lean and concrete. | 2 / 5 |
Actionability | Concrete, mostly copy-paste-ready commands are provided ("python scripts/hgnc.py fetch \"{term}\" --field \"{field}\"", "python -m py_compile scripts/hgnc.py") with example values ("BRAF", "HGNC:1097") in the parameter docs. It is not 5 because no fully worked invocation with real values and expected output is shown, leaving minor gaps. | 4 / 5 |
Workflow Clarity | The "Example run plan" lists a 4-step sequence, but checkpoints are vague ("Review the generated output and return the final artifact") and step 2 instructs editing an in-file "CONFIG block" that does not exist in scripts/hgnc.py. It is not 4 because validation is only implicit and one step is inaccurate; not 2 because the sequence is present and the operations (read-only API queries) are correctly outlined. | 3 / 5 |
Progressive Disclosure | A small, single-script skill with well-organized sections; the only bundle file referenced ("scripts/hgnc.py") exists and is correctly referenced throughout, and there is no content that warrants splitting into reference files. It is not 4 because navigation and content placement are appropriate for a skill of this size with no external references needed. | 5 / 5 |
Total | 14 / 20 Passed |