Content
88%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.
An exemplarily lean, actionable discipline skill: hard rules, a gated verification workflow with explicit failure loops, a claim-to-evidence mapping table, and a rigorous regression procedure with an executable patch-based revert. Its only notable weaknesses are the undefined `dev-smoke` invocation, the vague Task Tracing section, and no reference split for supplementary content.
Suggestions
Show how to run `dev-smoke` (an example command or pointer to its docs), since it is required evidence for any user-facing change.
Make Task Tracing concrete: give the actual probe/evidence-recording commands, or drop the section if tracing is environment-specific.
Consider moving the Red Flags and Rationalizations table to a one-level-deep reference file to keep the core gate workflow front and center.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body contains no explanations of concepts Claude already knows — even the git stash warning ("git stash pop restores whatever stash is on top, which may not be yours") is a non-obvious operational hazard, not padding — and every table row is a distinct directive, so every token earns its place. | 5 / 5 |
Actionability | Guidance is mostly executable: a numbered gate procedure, a claim-to-evidence table, a report template, an executable git diff/apply regression snippet, and a concrete memory command. However `dev-smoke` — the central user-facing verification mechanism — is never given an invocation command, and the Task Tracing section is vague ("run the real read probe first" with no command). | 4 / 5 |
Workflow Clarity | The Gate Function is a sequenced 0-5 process with an explicit failure loop ("If any step fails, stop. Fix the issue and restart from step 0"), and Regression Verification has must-pass/must-fail checkpoints plus a recovery rule ("If step 4 passes, the test is wrong. Rewrite it."), matching the top anchor for explicit validation and feedback loops. | 5 / 5 |
Progressive Disclosure | The skill is self-contained with clear, well-ordered section headers and easy navigation, but the ~91-line body exceeds the under-50-lines simple-skill exception and supplementary material (the Red Flags table, Task Tracing) is all inline with no one-level-deep reference split, leaving minor organization gaps. | 4 / 5 |
Total | 18 / 20 Passed |