Content
38%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 core usage documentation (parameters, data format, examples) is solid, but it is buried in generic process boilerplate, duplicated sections, dead references to nonexistent bundle files, and a wrong-domain audit command. The skill reads as template-padded rather than purpose-built for its forest-plot task.
Suggestions
Cut the boilerplate sections (Risk Assessment, Security Checklist, Lifecycle Status, Evaluation Criteria, Response Template, Output Contract, Validation and Safety Rules) — they are generic ceremony that consumes context without teaching the forest-plot task — and merge the duplicated Key Features/Features and Example Usage/Usage sections.
Remove or fix dead references: `references/` and `scripts/main.py` do not exist in this bundle, so either ship them or delete every section that points to them; also remove the nonsensical audit command that passes a clinical-note string as --input.
Consolidate the three overlapping workflow/run-plan sections into one validated sequence (compile check → run with confirmed inputs → review generated plot → fallback on failure) so the actual execution path is unambiguous.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is noticeably padded: boilerplate sections ("Risk Assessment", "Security Checklist", "Lifecycle Status" with a "Next Review Date", "Evaluation Criteria", "Response Template") plus duplicated content ("Key Features" vs "Features", "Example Usage" vs "Usage") and broken pointers ("See `## Features` above" points to a section that appears later). The core usage docs are buried in process ceremony. | 2 / 5 |
Actionability | The parameter table, input-data-format table, sample CSV, and example invocations are genuinely concrete, but the documented entry point `scripts/main.py` does not exist in the bundle and one "Audit-Ready Command" is nonsensical (passes a clinical-note string, "Audit validation sample with explicit symptoms, history...", as --input where a CSV path is required), leaving the guidance incomplete rather than copy-paste ready. | 3 / 5 |
Workflow Clarity | A sequence with some checkpoints exists (Quick Check `py_compile`, scope validation, error-handling fallbacks), but it is scattered across three overlapping sections ("Example run plan", "Workflow", "Error Handling") with no verify-then-fix loop on the generated output itself, matching the 'steps listed but checkpoints implicit/scattered' anchor. | 3 / 5 |
Progressive Disclosure | The body claims "Reference material available in `references/`" and "Reference guidance: `references/` contains supporting rules", but no references/ directory or any bundle file exists — dead references. Meanwhile content that belongs in separate files (preset styles, security checklist, risk assessment) is inlined into a ~350-line monolith. | 2 / 5 |
Total | 10 / 20 Passed |