Content
96%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 an excellent procedural skill: lean, fully actionable with concrete tool calls and field-level placement instructions, and a well-validated five-step workflow with explicit error-recovery branches and evidence-gating hard rules. The only structural note is that at its current length a small reference file could keep SKILL.md closer to overview size.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and directive throughout ('Call `list_materials` before drawing conclusions', 'Do not sample only the beginning') with no explanation of concepts Claude already knows and no padded sentences — every token earns its place, matching anchor 5. | 5 / 5 |
Actionability | Although instruction-only (no code, which is not penalized here), the guidance is fully executable: exact tool names, explicit branching on extraction state (idle/pending/failed), exact output field placement (`## 讲课风格` in `brief`, `generate_scene.materialFacts`), and copy-ready example values like 「以复习上节课的问题开场」. | 5 / 5 |
Workflow Clarity | Five clearly sequenced steps with explicit validation checkpoints and feedback loops: Step 1 verifies extraction completeness and handles each failure state, Step 4 verifies hypotheses with `search_material` and drops unsupported ones, and 'Hard rules' gates the final output — matching anchor 5. | 5 / 5 |
Progressive Disclosure | No bundle files exist and the ~77-line body is well organized into Step sections plus Hard rules, so nothing is misplaced; however it exceeds the under-50-line simple-skill guideline and the Step 2 style-profile checklist / example brief content are plausible candidates for a reference file, matching anchor 4 rather than 5. | 4 / 5 |
Total | 19 / 20 Passed |