Content
92%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 exemplary instruction-only skill body: lean, well-sectioned, with a mandatory gated sequence, explicit error-recovery rules, and a concrete handoff checklist. Its only gap is that some verification-oriented directives point at categories of commands rather than naming specific ones.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every line is a directive with no concept explanation, padding, or restatement of what Claude already knows (e.g. "Do not push when any gate is red", "Never edit an existing file in `migrations/"). Lean and efficient; each token earns its place. | 5 / 5 |
Actionability | Concrete executable commands are present (`just preflight`, `just gate-pr`), but several directives remain placeholders — "Targeted verification command", "Run targeted tests in addition to broad gate runs" — rather than specific runnable commands or examples. Not 5 because the guidance is not uniformly copy-paste ready; not 3 because the core flow commands are real and executable. | 4 / 5 |
Workflow Clarity | The "Mandatory Flow" is a numbered sequence with explicit validation checkpoints (gates red → no push) and a feedback loop for error recovery ("If the same command fails twice in one session, stop rerunning and switch to root-cause debugging"), plus a concrete recovery rule for migration flags. Not 4 because both the checkpoints and the retry/recovery loop are explicit. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines with no external material needed, and it is organized into clean one-glance sections (flow, review closure, async changes, migrations, handoff). Per the simple-skill guideline, well-organized sections with no reference indirection score 5; there are no buried or nested references. | 5 / 5 |
Total | 19 / 20 Passed |