Techwavedev Agi Agent Kit Figma Automation

Figma Automation via Rube MCP

Automate Figma operations through Composio's Figma toolkit via Rube MCP.

Prerequisites

• Rube MCP must be connected (RUBE_SEARCH_TOOLS available)
• Active Figma connection via RUBE_MANAGE_CONNECTIONS with toolkit figma
• Always call RUBE_SEARCH_TOOLS first to get current tool schemas

Setup

Get 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.

1. Verify Rube MCP is available by confirming RUBE_SEARCH_TOOLS responds
2. Call RUBE_MANAGE_CONNECTIONS with toolkit figma
3. If connection is not ACTIVE, follow the returned auth link to complete Figma auth
4. Confirm connection status shows ACTIVE before running any workflows

Core Workflows

1. Get File Data and Components

When to use: User wants to inspect Figma design files or extract component information

Tool sequence:

1. FIGMA_DISCOVER_FIGMA_RESOURCES - Extract IDs from Figma URLs [Prerequisite]
2. FIGMA_GET_FILE_JSON - Get file data (simplified by default) [Required]
3. FIGMA_GET_FILE_NODES - Get specific node data [Optional]
4. FIGMA_GET_FILE_COMPONENTS - List published components [Optional]
5. 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:

• Only supports Design files; FigJam boards and Slides return 400 errors
• ids must be a comma-separated string, not an array
• Node IDs may be dash-formatted (1-541) in URLs but need colon format (1:541) for API
• Broad ids/depth can trigger oversized payloads (413); narrow scope or reduce depth
• Response data may be in data_preview instead of data

2. Export and Render Images

When to use: User wants to export design assets as images

Tool sequence:

1. FIGMA_GET_FILE_JSON - Find node IDs to export [Prerequisite]
2. FIGMA_RENDER_IMAGES_OF_FILE_NODES - Render nodes as images [Required]
3. FIGMA_DOWNLOAD_FIGMA_IMAGES - Download rendered images [Optional]
4. FIGMA_GET_IMAGE_FILLS - Get image fill URLs [Optional]

Key parameters:

• file_key: File key
• ids: Comma-separated node IDs to render
• format: 'png', 'svg', 'jpg', or 'pdf'
• scale: Scale factor (0.01-4.0) for PNG/JPG
• images: Array of {node_id, file_name, format} for downloads

Pitfalls:

• Images return as node_id-to-URL map; some IDs may be null (failed renders)
• URLs are temporary (valid ~30 days)
• Images capped at 32 megapixels; larger requests auto-scaled down

3. Extract Design Tokens

When to use: User wants to extract design tokens for development

Tool sequence:

1. FIGMA_EXTRACT_DESIGN_TOKENS - Extract colors, typography, spacing [Required]
2. FIGMA_DESIGN_TOKENS_TO_TAILWIND - Convert to Tailwind config [Optional]

Key parameters:

• file_key: File key
• include_local_styles: Include local styles (default true)
• include_variables: Include Figma variables
• tokens: Full tokens object from extraction (for Tailwind conversion)

Pitfalls:

• Tailwind conversion requires the full tokens object including total_tokens and sources
• Do not strip fields from the extraction response before passing to conversion

4. Manage Comments and Versions

When to use: User wants to view or add comments, or inspect version history

Tool sequence:

1. FIGMA_GET_COMMENTS_IN_A_FILE - List all file comments [Optional]
2. FIGMA_ADD_A_COMMENT_TO_A_FILE - Add a comment [Optional]
3. FIGMA_GET_REACTIONS_FOR_A_COMMENT - Get comment reactions [Optional]
4. FIGMA_GET_VERSIONS_OF_A_FILE - Get version history [Optional]

Key parameters:

• file_key: File key
• as_md: Return comments in Markdown format
• message: Comment text
• comment_id: Comment ID for reactions

Pitfalls:

• Comments can be positioned on specific nodes using client_meta
• Reply comments cannot be nested (only one level of replies)

5. Browse Projects and Teams

When to use: User wants to list team projects or files

Tool sequence:

1. FIGMA_GET_PROJECTS_IN_A_TEAM - List team projects [Optional]
2. FIGMA_GET_FILES_IN_A_PROJECT - List project files [Optional]
3. 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 ID

Pitfalls:

• Team ID cannot be obtained programmatically; extract from Figma URL
• Only published styles/components are returned by team endpoints

Common Patterns

URL Parsing

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)

Node Traversal

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 details

Known Pitfalls

File Type Support:

• GET_FILE_JSON only supports Design files (figma.com/design/ or figma.com/file/)
• FigJam boards (figma.com/board/) and Slides (figma.com/slides/) are NOT supported

Node ID Formats:

• URLs use dash format: node-id=1-541
• API uses colon format: 1:541

Quick Reference

🧠 AGI Framework Integration

Adapted for @techwavedev/agi-agent-kit Original source: antigravity-awesome-skills

Hybrid Memory Integration (Qdrant + BM25)

Before executing complex tasks with this skill:

python3 execution/memory_manager.py auto --query "<task summary>"

Decision Tree:

• Cache hit? Use cached response directly — no need to re-process.
• Memory match? Inject context_chunks into your reasoning.
• No match? Proceed normally, then store results:

python3 execution/memory_manager.py store \
  --content "Description of what was decided/solved" \
  --type decision \
  --tags figma-automation <relevant-tags>

Note: Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.

Agent Team Collaboration

• Strategy: This skill communicates via the shared memory system.
• Orchestration: Invoked by orchestrator via intelligent routing.
• Context Sharing: Always read previous agent outputs from memory before starting.

Local LLM Support

When available, use local Ollama models for embedding and lightweight inference:

• Embeddings: nomic-embed-text via Qdrant memory system
• Lightweight analysis: Local models reduce API costs for repetitive patterns

Converted and distributed by TomeVault — claim your Tome and manage your conversions.