Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a well-structured, actionable workflow with explicit numbered steps, validation checkpoints, error-recovery loops, and properly signaled one-level-deep references to real bundle files. Its main weakness is mild verbosity — the 'Understanding Design Implementation' marketing-style section and some repeated design-token guidance restate things Claude already knows. Tightening those would lift conciseness to fully lean.
Suggestions
Remove or condense the 'Understanding Design Implementation' section ('For designers... For developers... For teams...') — it is promotional framing Claude already infers and adds no executable value.
De-duplicate the 'Avoid hardcoded values / prefer design system tokens' guidance, which appears nearly verbatim in Step 5, Step 6, Implementation Rules, and Common Issues; consolidate into one authoritative statement.
Trim 'Additional Resources' external links that duplicate the in-skill reference files, or move them into references/figma-tools-and-prompts.md to keep the body focused on the workflow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is largely efficient and actionable, but sections like 'Understanding Design Implementation' ('For designers: Confidence that implementations will match... For developers: A structured approach...') and repeated 'Avoid hardcoded values' guidance across Best Practices, Implementation Rules, and Step 6 pad the skill with content Claude already infers, keeping it just below fully lean. | 2 / 3 |
Actionability | Provides concrete, copy-paste-ready tool calls (`get_design_context(fileKey=":fileKey", nodeId="1-2")`, `codex mcp add figma --url ...`), a worked URL-parsing example with real values, and two end-to-end worked examples, all of which are directly executable. | 3 / 3 |
Workflow Clarity | The required workflow is a clearly numbered Step 0–7 sequence with an explicit validation checklist in Step 7, a validate→fix→retry loop for truncated responses ('If the response is too large... use get_metadata... fetch individual child nodes'), and clear error-recovery guidance in Common Issues. | 3 / 3 |
Progressive Disclosure | SKILL.md is an overview that signals one-level-deep references to two real bundle files ('references/figma-mcp-config.md', 'references/figma-tools-and-prompts.md'), both confirmed present, with setup/config and tool-catalog detail appropriately split out rather than inlined. | 3 / 3 |
Total | 11 / 12 Passed |