What problem does it solve? Documenting a design system component by hand means repeatedly switching between Figma and a spec template, and details like states, contrast ratios, and touch targets get missed or guessed. This Skill extracts visual properties directly from a Figma component and produces a complete, implementation-ready specification. ## Core Features & Use Cases - Figma-driven extraction: Uses MCP tools (get_design_context, get_screenshot, get_variable_defs) to pull layout, styles, variants, sizes, and design tokens from the component node. - 15-section spec output: Produces a spec covering anatomy, variants, sizes, states, accessibility, motion, haptics, platform differences, TypeScript API, usage examples, and anti-patterns, saved to docs/research/components/<tier>/<component-name>.md. - Provenance marking: Every observation is labeled confirmed (from Figma) or proposed (inferred from conventions), and ambiguities are logged to OPEN_QUESTIONS.md. - Use Case: Given a Figma button component URL and its atomic tier, generate a full spec with computed contrast ratios, 44pt touch target checks, and a draft props interface so a developer can implement it without opening Figma. ## Quick Start Generate a component spec for the Figma button component at this URL, tier atom, in my project workspace.