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 body delivers genuinely executable usage (parameters table, concrete CLI examples, output format, fallback paths) buried in heavy template boilerplate: self-referential sections, generic checklists, and vague pointer text to nonexistent detail. The core usable content is roughly a third of the file; the rest is padding that taxes the context window.
Suggestions
Cut the meta-sections that only point at other sections ("Key Features", "Dependencies", "Example Usage", "Implementation Details" openers like "See `## Features` above") and merge their one-line substance into the relevant sections.
Drop or drastically shrink the generic boilerplate (Risk Assessment, Security Checklist, Evaluation Criteria, Lifecycle Status, Response Template) — none of it is specific to comparison tables.
Replace the vague "`references/` contains supporting rules" pointers with a concrete link (e.g. "Table formatting standards: see [references/guidelines.md](references/guidelines.md)") and populate that file with the actual formatting standards it promises.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The 214-line body is noticeably verbose with padded template sections: "Key Features" and "Dependencies" sections that only say "See `## Features` above for related details", a "When to Use" bullet that repeats the frontmatter description verbatim ("task needs Auto-generates comparison tables..."), and generic boilerplate (Risk Assessment, Security Checklist, Evaluation Criteria, Lifecycle Status, Response Template) that adds almost no skill-specific information. Not 1 because it doesn't explain concepts Claude already knows — it's redundant scaffolding, not educational padding. | 2 / 5 |
Actionability | There is executable, copy-paste-ready guidance: a parameters table (`--items`, `--attributes`, `--output`), concrete invocations ("python scripts/main.py --items \"Drug A,Drug B\" --attributes \"Mechanism,Dose,Side Effects\""), a compile check, and a JSON output format. Not 5 because the "Example Usage" section's `cd "20260318/scientific-skills/..."` path is an odd hardcoded relative path, and no expected output of a real run is shown, leaving minor gaps. | 4 / 5 |
Workflow Clarity | The Workflow section gives a clear 5-step sequence (confirm objective → validate scope → execute with available inputs → structured result → fallback on failure) with checkpoints such as "stop early if the task would require unsupported assumptions" and Error Handling defines explicit failure paths. Not 5 because the validate→fix→retry feedback loop is described only abstractly — the sole concrete validation is `python -m py_compile scripts/main.py`, and the steps reference no per-step commands — and not 3 because the sequence and error/fallback checkpoints are genuinely present. | 4 / 5 |
Progressive Disclosure | Bundle structure is real (scripts/main.py and references/guidelines.md both exist), but references are only gestured at generically — "Reference material available in `references/`" and "`references/` contains supporting rules, prompts, or checklists" — with no named link to guidelines.md, and references/guidelines.md itself is just two bullet headings with no substance. Meanwhile large blocks of generic content (Security Checklist, Lifecycle Status, Evaluation Criteria) are inlined in SKILL.md where they add nothing, matching 'some structure but could be better organized; references present but not clearly signaled'. | 3 / 5 |
Total | 13 / 20 Passed |