Content
67%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 a well-structured, actionable instruction-only skill that gives Claude concrete commands, file paths, and example dialogues for an explicitly non-linear 'stance'. Its main weakness is token weight: the four large ASCII example dialogues and some redundant guardrails could be tightened or externalized.
Suggestions
Trim or consolidate the four ASCII-art entry-point examples (e.g., keep one representative dialogue and summarize the others) to reduce token weight and push conciseness up.
Deduplicate the 'Don't implement' guidance, which appears both in the intro IMPORTANT block and again in the Guardrails section.
Consider moving the lengthy entry-point example dialogues into a references/ file, referenced one level deep from the body, to bring progressive_disclosure to 5 and further cut SKILL.md tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient prose that assumes Claude's competence, but the four large ASCII-art entry-point examples and repeated guardrails ('Don't implement' appears in both the intro and Guardrails) consume meaningful tokens that could be trimmed, fitting 'mostly efficient but includes some unnecessary explanation or could be tightened'; it is not 2 because the content is not padded with concepts Claude already knows. | 3 / 5 |
Actionability | Provides concrete executable guidance throughout: a specific command ('openspec list --json'), exact artifact paths to read, a capture-location table, ready-made offer phrases, and full example dialogues for each entry point; it is not 5 because the example dialogues are illustrative rather than copy-paste-ready artifacts. | 4 / 5 |
Workflow Clarity | The 'When a change exists' branch lays out a clear 4-step numbered sequence with a capture table, and the read-only nature of the skill means no destructive-validation checkpoints are required (the destructive/batch cap does not apply); it is not 5 because the overall stance is deliberately non-linear, so explicit validation/feedback checkpoints are absent by design. | 4 / 5 |
Progressive Disclosure | A single well-organized SKILL.md with clear section headers (Stance, What You Might Do, OpenSpec Awareness, Entry Points, Ending, Guardrails) and no nested references; it is not 5 because the file is ~280 lines (well over the 50-line simple-skill threshold) and the long entry-point example dialogues could arguably live in a separate reference file. | 4 / 5 |
Total | 15 / 20 Passed |