Content
81%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 highly actionable and the workflows are explicitly sequenced with validation and feedback loops, which are its strongest dimensions. The main weakness is conciseness — several hard rules duplicate guidance already stated earlier in the file.
Suggestions
De-duplicate the Hard Rules against the lifecycle narrative: keep each rule in one place and cross-reference rather than restating 'never guess the port' and 'never spawn npm run dev' verbatim twice.
Consider moving the error-code table and state-files table into a references/ file (e.g. cli-reference.md) and signaling it from the body, to lift progressive_disclosure toward 5 and trim the inline token load.
Tighten the 'What this skill is NOT' and 'Lifecycle hooks and fallback' sections, which overlap with the 'Do NOT use' activation contract.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly dense and operational, but the ~350-line body restates the same rules in multiple places (e.g. 'never guess the port' and 'never spawn npm run dev' appear both in the lifecycle narrative and again in Hard Rules), so it could be tightened — matching the 3 anchor's 'mostly efficient but includes some unnecessary explanation or could be tightened' rather than the 4 anchor's 'minor instances'. | 3 / 5 |
Actionability | Fully executable guidance throughout — exact CLI verbs with flags, MCP tool signatures with parameters, concrete file paths, and an error-code recovery table — matching the 5 anchor's 'copy-paste ready commands; specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | The 'When the user just walked in' steps and the two-stage save→deploy→record→rollback flow are clearly sequenced with explicit validation (preview --status, JSON parsing, pre-flight env check) and a validate→fix→retry feedback loop, matching the 5 anchor; the destructive rollback path carries validation so the workflow cap does not apply. | 5 / 5 |
Progressive Disclosure | Well-organized with clear headers, tables, and one-level-deep external references signaled by skillName (ui-design, web-development, etc.) and the deployment-share.md pointer; the gap versus the 5 anchor is that the error-code table, state-files table, and detailed deploy workflow stay inline in a large SKILL.md rather than being split into reference files. | 4 / 5 |
Total | 17 / 20 Passed |