Content
53%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 guidance (CLI, parameters, examples, legend structure, well-split references) is concrete and executable, but the body is bloated with redundant governance boilerplate and the workflow lacks integrated validation checkpoints. Trimming the policy sections and sequencing verification into the workflow would materially raise quality.
Suggestions
Cut the generic governance sections (Risk Assessment, Security Checklist, Lifecycle Status, Evaluation Criteria, Required Inputs, Output Contract, User Checkpoints) and merge the three overlapping error-handling sections ('Workflow' step 5, 'Error Handling', 'Failure Handling') into one concise section — this alone would remove well over half the body.
Embed validation as sequenced workflow steps (e.g., step 0: 'python -m py_compile scripts/main.py'; after generation: verify the legend contains title, n, statistics, scale bar where applicable) instead of scattering it across 'Quick Check' and 'Quick Validation'.
Fix the broken third example ('python scripts/main.py --image.png --type scatter' is missing '--input') and either bundle a requirements.txt or replace the 'pip install -r requirements.txt' prerequisite with the actual dependency list (PIL, pytesseract).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Roughly half the body is generic skill-governance boilerplate (Risk Assessment, Security Checklist, Lifecycle Status, Evaluation Criteria, Required Inputs, Output Contract, User Checkpoints) with the same error-handling policy repeated across 'Workflow' step 5, 'Error Handling', and 'Failure Handling', plus a duplicated 'python scripts/main.py --help' command. It never explains concepts Claude already knows, so it is above 1, but it is well below the 'mostly efficient' bar of 3. | 2 / 5 |
Actionability | The Usage section provides an executable command, a complete parameter table, and concrete examples, and the Legend Structure plus real reference files give concrete output guidance. Not 5 because one example is broken ('python scripts/main.py --image.png --type scatter' is missing '--input') and Prerequisites references a requirements.txt that is not in the bundle. | 4 / 5 |
Workflow Clarity | The 5-step Workflow lists a sequence including an explicit fallback step, but validation checkpoints are scattered across disconnected sections ('Quick Check', 'Quick Validation') rather than sequenced into the workflow, matching the 'steps listed but checkpoints implicit' anchor. Not 4 because no workflow step embeds a verify-and-proceed gate like the 4-anchor example. | 3 / 5 |
Progressive Disclosure | The bundle split is genuinely good: legend templates and the academic style guide live in real one-level-deep reference files that are clearly listed in a References section, and the script is properly external. Not 5 because the SKILL.md overview itself carries 100+ lines of policy content that dilutes navigation and the references are listed at the end rather than signaled at point of use. | 4 / 5 |
Total | 13 / 20 Passed |