Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-engineered skill body: concise and table-driven, with concrete executable commands, clearly sequenced workflows that include verification, and a clean reference structure that points one level deep. It assumes Claude's competence and adds only package-specific knowledge.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and table/diagram-driven, assumes Claude's competence, and avoids explaining what Figma or parsers are; every section (architecture, key files, tasks, tests) earns its place without padding. | 3 / 3 |
Actionability | Provides copy-paste-ready, executable commands throughout — 'pnpm --filter @grida/io-figma fig2grida input.fig', 'npx tsx packages/.../fig2grida.ts input.fig --pages 0,2', the figma_archive.py invocation, and verification commands — plus specific file targets per task. | 3 / 3 |
Workflow Clarity | Multi-step tasks (add a property, debug clipboard paste, create fixtures) are clearly numbered and sequenced, the clipboard-debug flow has an explicit diff-against-schema feedback loop, and a dedicated Verification section supplies the typecheck/test checkpoint. | 3 / 3 |
Progressive Disclosure | A clear overview with a dedicated References table signaling one-level-deep pointers (.ref/figma/, the fig.kiwi glossary, the README) and cross-links to sibling skills (io-grida, render-reftest) for deeper detail, with content appropriately split into focused sections rather than inlined. | 3 / 3 |
Total | 12 / 12 Passed |