Content
52%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 is well-structured with real bundle files and a serviceable validated workflow, but it is padded with repetitive templated sections and its documented parameters diverge from the actual script. Tightening the boilerplate and reconciling the parameter docs would materially raise quality.
Suggestions
Reconcile the "Parameters" section with scripts/main.py's real flags (--name, --indication, --output, --demo) and add a concrete copy-paste run example using those flags.
Remove the duplicated description string from "When to Use"/"Key Features" and consolidate the redundant compile/help commands (Quick Check, Audit-Ready Commands, Example Usage) into one block.
Cut generic boilerplate (Risk Assessment table, Security Checklist, Evaluation Criteria, Lifecycle Status with a stale 2026-03-06 date) that does not advance the skill's actual task.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is noticeably verbose with many padded/templated sections (Risk Assessment, Security Checklist, Evaluation Criteria, Lifecycle Status) and repetition — the description string is restated verbatim in both "When to Use" and "Key Features", and `python -m py_compile scripts/main.py` appears in three sections; a stale "Next Review Date: 2026-03-06" also adds time-sensitive noise, fitting "noticeably verbose; several unnecessary padded sections." | 2 / 5 |
Actionability | Concrete executable commands exist (`python -m py_compile scripts/main.py`, `python scripts/main.py --help`) and a real script is referenced, but the documented "Parameters" (drug_profile, comparator, market) do not match the script's actual flags (--name, --indication, --output, --demo) and no copy-paste run invocation is given, fitting "some concrete guidance but incomplete; missing key details." | 3 / 5 |
Workflow Clarity | The Workflow section gives a clear 5-step sequence with an explicit stop-early validation (step 2) and a documented fallback path (step 5), plus a dedicated Error Handling section, fitting "clear sequence with most checkpoints present"; it does not reach 5 because there is no output-validation feedback loop (validate generated artifact -> fix -> retry). | 4 / 5 |
Progressive Disclosure | Bundle files are real and referenced one level deep with clear signaling ([references/audit-reference.md], scripts/main.py), and the body is organized into titled sections, fitting "good structure; references mostly clear"; it stays at 4 rather than 5 because the SKILL.md overview itself is bloated with generic sections that could be trimmed or split out. | 4 / 5 |
Total | 13 / 20 Passed |