Content
87%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.
A concise, highly actionable body that routes the agent to real helper scripts with concrete flags and examples, and is well-structured with verified one-level-deep references. The main weakness is workflow_clarity: batch install operations lack explicit validation/retry checkpoints, which the rubric caps at 3.
Suggestions
Add an explicit validation checkpoint after install (e.g. verify the skill directory now exists and contains a SKILL.md, and report failures with a retry path) to lift workflow_clarity past the destructive/batch cap of 3.
Document the error-recovery flow when a download fails with auth errors before falling back to git sparse checkout (current state vs. next action) so the install loop has a clear feedback step.
For multi-skill installs, note a per-skill success/failure summary so batch operations surface partial failures explicitly.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean body that assumes Claude's competence with no padding or concept over-explanation; each section (Communication, Scripts, Behavior and Options, Notes) earns its place and delivers only skill-specific facts. | 5 / 5 |
Actionability | Provides concrete, copy-paste-ready script invocations with real flags and examples (e.g. 'scripts/install-skill-from-github.py --repo <owner>/<repo> --path <path/to/skill>', '--format json', experimental examples) covering the common cases. | 5 / 5 |
Workflow Clarity | Decision routing (list vs install-by-name vs install-by-url) and behavior guards are present, but installation is a batch/destructive-class operation with no explicit validate→fix→retry feedback loop; the rubric caps such skills at 3 when validation checkpoints are absent. | 3 / 5 |
Progressive Disclosure | Well-organized sections with all referenced script paths existing one level deep in scripts/ (list-skills.py, install-skill-from-github.py); a short single-purpose skill with no nested references, so structure fully satisfies navigation needs. | 5 / 5 |
Total | 18 / 20 Passed |