Content
85%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.
A highly actionable, well-structured skill with excellent workflow sequencing and progressive disclosure. The main weakness is conciseness — the verbatim-relay and mock-prefix guidance is restated across several sections and could be consolidated.
Suggestions
Consolidate the repeated "courier, not co-author" and X-MOCK-prefix guidance into a single authoritative section; cross-reference it instead of restating the rule in the intro, Output contract, hook subsection, and self-check.
Trim the Gotchas entries to the concrete flag/env facts (e.g. CLAUDECODE="", OpenCode provider/model form, cursor --mode) and drop the surrounding rationale paragraphs that re-explain why.
Reduce the self-check section to a compact 3-item checklist rather than the multi-paragraph right/wrong examples that largely restate the Output contract.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient but several sections are padded or repeated — the courier rule and "X-MOCK prefix" guidance are stated multiple times, and the self-check/gotchas sections run long with restated rationale that Claude already grasps. | 3 / 5 |
Actionability | Fully executable guidance: copy-paste parallel-run recipe, per-CLI binary/flag/model-form cheatsheet, and concrete discovery commands covering the common cases. | 5 / 5 |
Workflow Clarity | Clear sequenced workflow (parse → pick model → build & run → present) with an explicit self-check validation checkpoint and error-recovery guidance for hook-blocked and failed runs. | 5 / 5 |
Progressive Disclosure | SKILL.md is a clear overview with two well-signaled one-level-deep references (references/models.md, references/runtimes.md) that hold the bulk detail; navigation is easy and no nesting beyond one level. | 5 / 5 |
Total | 18 / 20 Passed |