Content
100%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 lean skill body: dense decision rules, a ready-to-dispatch reviewer prompt, explicit verdict and retry semantics, and a complete escalation table. It fully respects Claude's competence while adding exactly the procedural detail Claude could not infer.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every sentence is a rule or decision input — skip conditions, reviewer inputs, verdict criteria, escalation table — with zero explanation of concepts Claude already knows. The 44-line body earns every token; even rationale sentences like 'A panel costs three reviews' carry a decision rule. | 5 / 5 |
Actionability | Provides a fully copy-paste-ready reviewer prompt template with explicit placeholders (<title>, <diff>, <its feedback>), the exact inputs to supply and withhold, a concrete verdict format with severity levels, and specific retry limits ('retry N of 2'). An executable example covers the common case end to end. | 5 / 5 |
Workflow Clarity | Clear sequenced workflow with explicit validation checkpoints: run tests/lint/build first, gate acceptance on the PASS/FAIL verdict, retry with feedback passed forward ('Previous FAIL, on a retry: <its feedback>'), and a defined escalation path (panel of three, then stop and tell the user). The feedback loop for the FAIL-retry cycle is fully specified. | 5 / 5 |
Progressive Disclosure | A single-purpose skill under 50 lines with no need for external references: content is split into well-organized, clearly headed sections (When to skip it, Review, After a FAIL) and stays one level deep. Meets the simple-skill exception for a top structure score. | 5 / 5 |
Total | 20 / 20 Passed |