Content
86%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.
The body is a tight, well-sequenced review playbook with concrete steps, a verification checkpoint, and clear synthesis categories; it stays lean and needs no external references, with only minor actionability and feedback-loop gaps.
Suggestions
Add an explicit feedback loop for step 5, e.g. "If a load-bearing claim fails verification, demote it to a follow-up and re-synthesize before recording decisions."
Give one concrete line on how to choose the "representative users plus specialists" in step 3 so reviewer selection is repeatable rather than left to judgment.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is a lean seven-step list plus two short guardrail lines with no padding and no explanation of concepts Claude already knows; every line earns its place. | 5 / 5 |
Actionability | Steps are concrete and directive ("Build one evidence package from current source, live behavior, and relevant public contracts", "Synthesize consensus under: Fix before change / Fix now / Follow-up / No action", "Record approved decisions in the active `.plans/<task>.md`"), with only minor gaps such as how to actually select or recruit reviewers. | 4 / 5 |
Workflow Clarity | A clear numbered seven-step sequence includes an explicit verification checkpoint ("Verify load-bearing claims against source before accepting them") and a synthesis framework, but there is no explicit error-recovery feedback loop after verification fails. | 4 / 5 |
Progressive Disclosure | The skill is under 50 lines, needs no external references, and is well organized as a numbered process with closing guardrails, satisfying the simple-skill exception for progressive disclosure. | 5 / 5 |
Total | 18 / 20 Passed |