Design Lazyyy — Figma MCP Server for AI-powered design system analysis, token extraction, and read/write operations
Server has 19 tools with consistent naming (verb_noun pattern) and descriptions present for all. However, parameter descriptions are sparse or generic, input schemas lack depth (missing min/max constraints, format specs), and output schemas are not documented. Error handling returns basic text responses without recovery guidance. Tool composition is sound but lacks idempotency hints and chaining metadata. Descriptions average ~100 chars (below 194-char baseline), and many parameters lack type constraints beyond basic JSON Schema.
Compare design token values from Figma against code token values to find discrepancies. Returns a parity score (0-100) with categorized differences.
Clear all captured console logs. Only available in local mode.
Delete a comment from a Figma file by comment ID.
Get all comments from a Figma file. Supports markdown output with threaded replies.
Get component metadata and reconstruction spec from a Figma file. Search by node ID or name. Returns layout, styles, and properties for code generation.
Generate comprehensive component documentation from a Figma component including anatomy, variants, states, design tokens, accessibility notes, and handoff checklist.
Output schemas not documented. Tools return JSON/text responses but LLMs cannot infer field structure, types, or required fields for downstream chaining. E.g., figma_get_file_data returns a document tree but no schema is visible.
Parameter descriptions lack format/constraint details. E.g., 'depth' in figma_get_file_data has no min/max; 'format' enums exist but no guidance on when to use each; 'scale' in figma_get_component_image has no range (0.1 - 10? 1 - 100?).
Error handling returns plain text 'Error: <message>' without recovery guidance. LLMs cannot determine if errors are retryable, user-fixable, or fatal. E.g., 'Error: Invalid file key' gives no hint to call figma_get_status or search for the file.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 53 | <=2025-11-25 | v2 |
Render Figma nodes as images (PNG, SVG, PDF, JPG). Returns image URLs for the specified node IDs.
Get console logs captured from the Figma plugin context. Only available in local mode with Desktop Bridge.
Extract a unified design system kit from a Figma file: tokens, components, styles, relationships, and quality scores. Uses UX/UI Architect intelligence to score across 6 dimensions: Visual Hierarchy, Consistency, Accessibility, Usability, Responsiveness, Performance.
Retrieves file data from Figma including document structure, components, and styles. Accepts a Figma file URL or file key, with optional node ID and depth parameters.
Get metadata about a Figma file without fetching the full document tree. Returns name, last modified, version, components, styles, and branches.
Get version history of a Figma file.
Check connection status by fetching the authenticated user's Figma profile. Also shows Desktop Bridge status for write operations.
Extract all styles (colors, text, effects, grids) from a Figma file with enriched properties and multi-format code export (CSS, SCSS, Tailwind, JSON, TypeScript).
Extract design variables/tokens from a Figma file with alias resolution and multi-format code export. Supports CSS, SCSS, Tailwind, TypeScript, and JSON output.
Lint a Figma design for accessibility (WCAG 2.2 AA), consistency, and design system compliance. Checks contrast ratios, target sizes, heading hierarchy, spacing grid, and token usage. Returns categorized findings: Critical > Major > Minor > Enhancement.
Generate a direct URL to a specific node in a Figma file for quick navigation.
Post a comment on a Figma file. Can be attached to a node, pinned to coordinates, or a reply to another comment.
Get a snapshot of recent console activity and bridge status. Use repeatedly to 'watch' console output. Only available in local mode.
No idempotency hints on write tools. figma_post_comment and figma_delete_comment lack guidance on retry safety. If an agent retries after a network timeout, will it create duplicate comments or fail safely?
Tool descriptions average ~100 chars; baseline is 194 chars. Many lack WHEN-to-use context. E.g., 'Get version history' does not explain when to call it vs figma_get_file_metadata, or what the response structure is.