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 orchestration skill body: a tightly sequenced, evidence-gated workflow with concrete commands, explicit feedback loops, and clean progressive disclosure into two real reference files. The only nit is a few justificatory 'Why:' passages that slightly pad the token budget.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and purposeful with no basic-concept padding and clear role framing, but the inline 'Why:' rationale paragraphs (e.g., on evidence and concurrency) are minor over-explanation that could be trimmed while keeping the actionable core. | 4 / 5 |
Actionability | Concrete directives throughout — '/code-review --fix' (effort high/xhigh), exact model assignments (sonnet/haiku/session), real file references, and invocation examples — with exact role recipes delegated to existing reference files, giving fully executable guidance. | 5 / 5 |
Workflow Clarity | Explicit fixed sequence (Step −1 → freeze → scouts → fan-out → per-phase loop → integration → whole-diff → acceptance) with evidence-based NON-NEGOTIABLE GATES, feedback loops on FAIL, and a DONE checklist. | 5 / 5 |
Progressive Disclosure | Overview body with well-signaled one-level-deep references to real files ('references/role-prompts.md', 'references/templates.md'), with role-prompt and template detail appropriately split out and clearly signaled for navigation. | 5 / 5 |
Total | 19 / 20 Passed |