Content
96%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 an exemplar of an actionable runtime-QA skill: dense, executable, and assumption-respecting, with a well-sequenced workflow, explicit validation, and useful checklists. The only soft spot is progressive disclosure — references are repo-relative and well-signaled but there is no structured reference bundle to verify against.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and high-signal throughout — it assumes Claude's competence and gives only non-obvious domain rationale (Apple Silicon code-signature cache poisoning, the stale-PATH false pass) rather than explaining basics, so every token earns its place. | 5 / 5 |
Actionability | Fully executable, copy-paste-ready commands cover the common cases: the stamped build, install-dogfood.sh, fresh-shell verification, and the app-server/exec smoke — each with exact env vars and flags. | 5 / 5 |
Workflow Clarity | A clear 7-step sequence with explicit validation checkpoints (gating, embedded-SHA verification, fresh-shell resolution, version-string check), false-pass warnings, and 'Red flags / don't' plus 'Output' sections that function as checklists for a destructive install. | 5 / 5 |
Progressive Disclosure | Well-organized into When-to-use / Workflow / Red-flags / Output sections with clearly signaled one-level-deep references (crates/tui/AGENTS.md, docs/MOTION_CONTRACT.md, sibling cw-* skills); the inlined scenario checklist is appropriate, but no dedicated reference bundle exists and a little scenario detail could live separately, so it sits just below the exemplar. | 4 / 5 |
Total | 19 / 20 Passed |