Content
100%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-organized, instruction-style skill: an explicit ordered workflow with a validation checkpoint, concrete tool calls and rules, and clean one-level references that exist on disk. It assumes Claude's intelligence and stays tight. Strong result with no material weaknesses.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean, assumes Claude's competence (no explanation of what Figma/MCP is), and every section carries concrete rules; only minor throat-clearing in the intro/meta sentence keeps it just at the efficient top anchor. | 3 / 3 |
Actionability | Gives concrete, executable guidance naming specific MCP tools (get_design_context, get_metadata, get_screenshot) and explicit rules ("DO NOT import/add new icon packages", "Replace Tailwind utility classes with the project's preferred tokens"); absence of raw code is fine for this instruction-only skill. | 3 / 3 |
Workflow Clarity | A clearly numbered "Required flow (do not skip)" with a truncation fallback loop (get_metadata then re-fetch) and a validation checkpoint ("Validate against Figma for 1:1 look and behavior before marking complete"), matching the explicit-validation-with-feedback anchor. | 3 / 3 |
Progressive Disclosure | A concise overview signals two real one-level-deep references (figma-mcp-config.md and figma-tools-and-prompts.md), both present on disk and restated in a References section, with no nested/inline walls of text. | 3 / 3 |
Total | 12 / 12 Passed |