Content
93%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, actionable skill body: concrete API payloads, response contracts, and a sequenced recipe with a fallback path, all with zero padding. The only refinement would be a post-apply verification step to close the workflow's feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient: the body covers auth ('OD_TOOL_TOKEN, injected by the daemon'), exact endpoint payloads, response shapes, the 'semantic: false' caveat, and the fallback rule without a single concept Claude already knows or any padding — every token earns its place, matching the anchor-5 example of terse section-plus-code structure. | 5 / 5 |
Actionability | Fully executable guidance: copy-paste-ready JSON bodies for both endpoints ('POST /api/tools/library/search' and 'POST /api/tools/library/apply'), concrete response shapes to branch on, a wiring example ('<img src="assets/ab12cd34ef.png">'), and an explicit guardrail ('never invent a relPath that apply did not return') covering the common cases. | 5 / 5 |
Workflow Clarity | The Recipe section gives a clear sequenced process (search -> pick best result by dimensions/title/source -> apply -> wire the path in) with an error-recovery branch ('If search returns nothing, fall back to media generation'), matching anchor 4. It falls short of anchor 5 because there is no explicit verification checkpoint after apply (e.g., confirming the asset landed at relPath before wiring it into the artifact). | 4 / 5 |
Progressive Disclosure | This is a simple single-purpose skill (~50 lines, no external references needed) with well-organized sections (When to use / Tools / Recipe), and no bundle files exist to navigate — per the simple-skill exception this is a 5; nothing that belongs in a separate file is inlined. | 5 / 5 |
Total | 19 / 20 Passed |