Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Automate Figma tasks via Rube MCP (Composio): files, components, design tokens, comments, exports. Always search tools first for current schemas.
.claude/skills/davepoon-figma-automation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 118% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 140% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 84% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 166% | 0% |
Automate Figma operations through Composio's Figma toolkit via Rube MCP.
Toolkit docs: composio.dev/toolkits/figma
RUBE_MANAGE_CONNECTIONS with toolkit figmaRUBE_SEARCH_TOOLS first to get current tool schemasGet Rube MCP: Add https://rube.app/mcp as an MCP server in your client configuration. No API keys needed — just add the endpoint and it works.
RUBE_SEARCH_TOOLS respondsRUBE_MANAGE_CONNECTIONS with toolkit figmaWhen to use: User wants to inspect Figma design files or extract component information
Tool sequence:
FIGMA_DISCOVER_FIGMA_RESOURCES - Extract IDs from Figma URLs Prerequisite]FIGMA_GET_FILE_JSON - Get file data (simplified by default) Required]FIGMA_GET_FILE_NODES - Get specific node data Optional]FIGMA_GET_FILE_COMPONENTS - List published components Optional]FIGMA_GET_FILE_COMPONENT_SETS - List component sets Optional]Key parameters:
file_key: File key from URL (e.g., 'abc123XYZ' from figma.com/design/abc123XYZ/...)ids: Comma-separated node IDs (NOT an array)depth: Tree traversal depth (2 for pages and top-level children)simplify: True for AI-friendly format (70%+ size reduction)Pitfalls:
ids must be a comma-separated string, not an arraydata_preview instead of dataWhen to use: User wants to export design assets as images
Tool sequence:
FIGMA_GET_FILE_JSON - Find node IDs to export Prerequisite]FIGMA_RENDER_IMAGES_OF_FILE_NODES - Render nodes as images Required]FIGMA_DOWNLOAD_FIGMA_IMAGES - Download rendered images Optional]FIGMA_GET_IMAGE_FILLS - Get image fill URLs Optional]Key parameters:
file_key: File keyids: Comma-separated node IDs to renderformat: 'png', 'svg', 'jpg', or 'pdf'scale: Scale factor (0.01-4.0) for PNG/JPGimages: Array of {node_id, file_name, format} for downloadsPitfalls:
When to use: User wants to extract design tokens for development
Tool sequence:
FIGMA_EXTRACT_DESIGN_TOKENS - Extract colors, typography, spacing Required]FIGMA_DESIGN_TOKENS_TO_TAILWIND - Convert to Tailwind config Optional]Key parameters:
file_key: File keyinclude_local_styles: Include local styles (default true)include_variables: Include Figma variablestokens: Full tokens object from extraction (for Tailwind conversion)Pitfalls:
When to use: User wants to view or add comments, or inspect version history
Tool sequence:
FIGMA_GET_COMMENTS_IN_A_FILE - List all file comments Optional]FIGMA_ADD_A_COMMENT_TO_A_FILE - Add a comment Optional]FIGMA_GET_REACTIONS_FOR_A_COMMENT - Get comment reactions Optional]FIGMA_GET_VERSIONS_OF_A_FILE - Get version history Optional]Key parameters:
file_key: File keyas_md: Return comments in Markdown formatmessage: Comment textcomment_id: Comment ID for reactionsPitfalls:
When to use: User wants to list team projects or files
Tool sequence:
FIGMA_GET_PROJECTS_IN_A_TEAM - List team projects Optional]FIGMA_GET_FILES_IN_A_PROJECT - List project files Optional]FIGMA_GET_TEAM_STYLES - List team published styles Optional]Key parameters:
team_id: Team ID from URL (figma.com/files/team/TEAM_ID/...)project_id: Project IDPitfalls:
Extract IDs from Figma URLs:
1. Call FIGMA_DISCOVER_FIGMA_RESOURCES with figma_url
2. Extract file_key, node_id, team_id from response
3. Convert dash-format node IDs (1-541) to colon format (1:541)1. Call FIGMA_GET_FILE_JSON with depth=2 for overview
2. Identify target nodes from the response
3. Call again with specific ids and higher depth for detailsFile Type Support:
Node ID Formats:
node-id=1-5411:541| Task | Tool Slug | Key Params | |------|-----------|------------| | Parse URL | FIGMA_DISCOVER_FIGMA_RESOURCES | figma_url | | Get file JSON | FIGMA_GET_FILE_JSON | file_key, ids, depth | | Get nodes | FIGMA_GET_FILE_NODES | file_key, ids | | Render images | FIGMA_RENDER_IMAGES_OF_FILE_NODES | file_key, ids, format | | Download images | FIGMA_DOWNLOAD_FIGMA_IMAGES | file_key, images | | Get component | FIGMA_GET_COMPONENT | file_key, node_id | | File components | FIGMA_GET_FILE_COMPONENTS | file_key | | Component sets | FIGMA_GET_FILE_COMPONENT_SETS | file_key | | Design tokens | FIGMA_EXTRACT_DESIGN_TOKENS | file_key | | Tokens to Tailwind | FIGMA_DESIGN_TOKENS_TO_TAILWIND | tokens | | File comments | FIGMA_GET_COMMENTS_IN_A_FILE | file_key | | Add comment | FIGMA_ADD_A_COMMENT_TO_A_FILE | file_key, message | | File versions | FIGMA_GET_VERSIONS_OF_A_FILE | file_key | | Team projects | FIGMA_GET_PROJECTS_IN_A_TEAM | team_id | | Project files | FIGMA_GET_FILES_IN_A_PROJECT | project_id | | Team styles | FIGMA_GET_TEAM_STYLES | team_id | | File styles | FIGMA_GET_FILE_STYLES | file_key | | Image fills | FIGMA_GET_IMAGE_FILLS | file_key |
Powered by Composio
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 8,018 | 7,185 | -10% | 1 | 1 | 0% | 1,256 | 2,740 | +118% | 0 | 0 | — |
case-01 | fail→fail | 8,702 | 5,552 | -36% | 1 | 1 | 0% | 1,782 | 2,420 | +36% | 0 | 0 | — |
case-02 | fail→fail | 17,753 | 5,515 | -69% | 1 | 1 | 0% | 3,979 | 2,362 | -41% | 0 | 0 | — |
case-03 | fail→fail | 10,765 | 5,877 | -45% | 1 | 1 | 0% | 1,147 | 2,313 | +102% | 0 | 0 | — |
case-04 | fail→fail | 4,499 | 7,987 | +78% | 1 | 1 | 0% | 818 | 2,502 | +206% | 0 | 0 | — |
case-05 | fail→fail | 8,819 | 6,594 | -25% | 1 | 1 | 0% | 420 | 2,557 | +509% | 0 | 0 | — |
case-06 | fail→fail | 10,433 | 4,846 | -54% | 1 | 1 | 0% | 1,092 | 2,202 | +102% | 0 | 0 | — |
case-08 | pass→pass | 9,702 | 4,539 | -53% | 1 | 1 | 0% | 1,593 | 2,655 | +67% | 0 | 0 | — |
case-09 | pass→pass | 3,624 | 3,177 | -12% | 1 | 1 | 0% | 588 | 2,626 | +347% | 0 | 0 | — |
case-10 | fail→fail | 6,571 | 3,014 | -54% | 1 | 1 | 0% | 1,105 | 2,465 | +123% | 0 | 0 | — |
case-11 | fail→pass | 6,401 | 8,732 | +36% | 1 | 1 | 0% | 1,261 | 3,024 | +140% | 0 | 0 | — |
case-12 | pass→pass | 3,924 | 2,329 | -41% | 1 | 1 | 0% | 728 | 2,436 | +235% | 0 | 0 | — |
case-13 | pass→pass | 8,176 | 6,399 | -22% | 1 | 1 | 0% | 1,403 | 3,065 | +118% | 0 | 0 | — |
case-14 | pass→pass | 3,395 | 3,545 | +4% | 1 | 1 | 0% | 697 | 2,682 | +285% | 0 | 0 | — |
case-15 | fail→fail | 4,859 | 7,928 | +63% | 1 | 1 | 0% | 873 | 2,870 | +229% | 0 | 0 | — |
case-16 | fail→pass | 13,088 | 4,688 | -64% | 1 | 1 | 0% | 2,367 | 2,813 | +19% | 0 | 0 | — |
case-17 | fail→pass | 8,277 | 4,835 | -42% | 1 | 1 | 0% | 1,613 | 2,973 | +84% | 0 | 0 | — |
case-18 | fail→fail | 5,093 | 6,038 | +19% | 1 | 1 | 0% | 991 | 2,341 | +136% | 0 | 0 | — |
case-19 | fail→fail | 3,897 | 4,927 | +26% | 1 | 1 | 0% | 632 | 2,270 | +259% | 0 | 0 | — |
case-20 | fail→pass | 5,183 | 3,247 | -37% | 1 | 1 | 0% | 960 | 2,556 | +166% | 0 | 0 | — |
case-21 | fail→fail | 5,839 | 5,166 | -12% | 1 | 1 | 0% | 1,005 | 2,378 | +137% | 0 | 0 | — |
case-22 | fail→fail | 4,929 | 7,031 | +43% | 1 | 1 | 0% | 963 | 2,299 | +139% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 11 counted toward the lift figure. The other 11 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 11 comparable cases. 6 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.