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.
A tight, well-structured instruction skill that gives concrete per-runtime tool guidance with clear decision criteria and boundary notes. The only gap is the absence of a complete example card/invocation for the managed-agent path, which keeps actionability at 4 rather than 5.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — the opening 'Escalate to a real person when guessing is expensive' and the dense runtime sections carry only actionable specifics (required fields, icon rules, constraints) with no padding or explanation of concepts Claude already knows, matching the 'lean and efficient; every token earns its place' anchor. | 5 / 5 |
Actionability | Provides concrete, executable guidance — exact tool names, exact params (kind, card.title, card.options 2–10), exact CLI commands (human ask|approve|choose|tell) — but the managed-agent case never shows a complete example card/invocation shape, leaving a minor gap versus the 'copy-paste ready code covering common cases' 5 anchor. | 4 / 5 |
Workflow Clarity | This is a simple decision-plus-single-action skill: the 'Use it when' criteria, the per-kind guidance (choose for multi-way, tell for notify), and the three unambiguous runtime branches make the action clear, satisfying the simple-skill exception for a 5; the destructive-operation validation cap does not apply because HITL is itself the validation gate. | 5 / 5 |
Progressive Disclosure | Under 50 lines, single-purpose, with no need for external bundle files; the content is organized into a clear intro, a 'Use it when' list, boundary notes, and three runtime subsections, matching the 'well-organized sections' exception for a 5. | 5 / 5 |
Total | 19 / 20 Passed |