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.
Excellent operational content: executable code with built-in verification loops, explicit rules for destructive and ambiguous operations, and dense skill-specific guidance with zero padding. The only structural note is that everything lives in a single file when a small reference file could absorb the peripheral API detail.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and imperative throughout ("Default to multiplayer-safe, surgical `editContent`", "Base64 is only for image data URLs inside Markdown") and explains only Figma-spec-specific API behavior Claude would not know; every section earns its tokens. | 5 / 5 |
Actionability | Fully executable, copy-paste-ready code covers creation (`figma.spec.createOnCanvasAsync`), all three content methods (`getContent`/`setContent`/`editContent`), and comments (`figma.spec.comments.*`), with exact call signatures like `findAllWithCriteria({ types: ['SPEC_BLOCK'] })` in the prose. | 5 / 5 |
Workflow Clarity | Workflows include explicit validation and feedback loops: `getContent` before edits, post-write checks ("if (!after.markdown.includes('Edited first paragraph.')) throw"), error-recovery guidance ("Add surrounding context to make it unique"), and a five-item Critical rules checklist; the destructive `setContent` is gated to "an empty first draft". | 5 / 5 |
Progressive Disclosure | Sections are clean and navigable (Discover and choose, Create, Content methods, Critical rules, Style, Images, Comments) with no nested references, but at ~105 lines — above the rubric's under-50-line simple-skill exception — all reference material (e.g., full comments/images API detail) is inlined rather than split into bundle files. | 4 / 5 |
Total | 19 / 20 Passed |