figma

name	figma
description	Access Figma design files using native pi tools — read LLM-ready summaries, explanations, implementation context, screenshots, components, styles, variables, and design tokens. Requires a Figma personal access token.

Figma Design Integration

Use the native figma_* tools to read Figma files and translate designs into code. Prefer processed, LLM-ready tools over raw Figma JSON.

When to Use

• User provides a Figma file URL and asks you to explain or implement a design.
• User asks about Figma colors, typography, spacing, components, variables, or layout.
• You need screenshots/assets for visual validation.

Authentication

The tools read the token from an in-memory override, FIGMA_TOKEN, or ~/.pi/agent/auth.json at .figma.token.

If auth is missing, invalid, or expired, do not ask the user to paste the token in chat. Use the native figma_configure_auth tool or ask the user to run /figma-auth --force. The prompt is masked and the token is stored by the extension without returning it to the model.

URL Parsing

Given:

https://www.figma.com/design/ABC123def456/Project-Name?node-id=123-456

• File key: ABC123def456
• Node ID: 123-456 from the URL (figma_* tools also accept API format 123:456)

Use figma_parse_url when you need to extract these values from a full URL.

Default Tool Workflow

1. Use figma_configure_auth only when auth is missing, invalid, expired, or the user asks to update the token.
2. Use figma_parse_url for full Figma URLs.
3. If the exact target node is unclear, use figma_get_design_context plus figma_find_nodes_by_name or figma_find_nodes_by_text before raw exploration.
4. Use figma_render_nodes for screenshots and figma_extract_assets when icons, node renders, or image fills are needed for implementation.
   • Do not pass outputDir unless the user explicitly requests persistent files in a specific location. Let the extension use its OS temp directory default for generated images/assets.
5. Use figma_explain_node or figma_get_node_summary for the target frame/component.
6. Use enriched figma_get_implementation_context when coding from a design. Pass framework, styling, resolveTokens, and includeCodeSnippets when useful.
7. Use figma_find_code_connect_mapping only when the local repo may contain Code Connect mappings or Figma URL/node references.
8. Use figma_get_component_implementation_hints for component-level implementation planning.
9. Use figma_get_nodes only for raw debugging.

Do not call figma_get_nodes by default. Prefer processed tools.

Prefer batch calls where supported: pass multiple node IDs to figma_render_nodes, figma_get_node_metadata, or raw figma_get_nodes instead of looping.

Recommended Workflows

Explaining a component

figma_parse_url
figma_render_nodes
figma_explain_node

Implementing a design

figma_parse_url
figma_find_nodes_by_name/text if the URL lacks an exact target node
figma_render_nodes
figma_get_implementation_context with framework/styling/token options when useful
figma_get_node_summary for specific subnodes if needed

Extracting assets

figma_extract_assets

Use this when visual assets are needed. Prefer the manifest's nodePath, suggestedName, hashes, and local paths over guessing layer-to-file mappings.

Finding local implementation mappings

figma_find_code_connect_mapping
figma_get_component_implementation_hints

Only use these when you are working in a local code repository where Code Connect or Figma URL/node references may exist.

Debugging the extension or raw Figma data

figma_get_nodes

Processed Tools

• figma_get_node_summary returns compact structured summaries: name, type, size, layout, spacing/padding, fills/strokes/effects, visible text, component properties, and immediate child hierarchy.
• figma_extract_text returns visible text nodes only.
• figma_explain_node returns human-readable Markdown for questions like “Explain this component.”
• figma_find_nodes_by_name and figma_find_nodes_by_text search names/text with compact path-aware matches. Use them before raw file exploration.
• figma_get_implementation_context returns coding-ready context: purpose, sections, fields/buttons, measurements, typography, colors, spacing, CSS layout/responsive hints, accessibility hints, design tokens, assets, framework hints, and hierarchy.
• figma_extract_assets returns an asset manifest for SVG icons, node renders, and image fills.
• figma_find_code_connect_mapping scans the local repo for Code Connect/Figma references.
• figma_get_component_implementation_hints combines Figma context, variants, tokens, assets, accessibility, Code Connect matches, and optional starter snippets.

Processed tools default to shallow, safe fetches:

{
  depth: 2,
  includeHidden: false,
  includeVectors: false,
  includeComponentInternals: false,
  framework?: "react" | "html" | "vue" | "angular" | "react-native",
  styling?: "css" | "css-modules" | "styled-components" | "tailwind" | "inline",
  resolveTokens?: true,
  includeCodeSnippets?: false
}

Only increase depth (max 4) or enable internals/vectors for a specific child node when needed.

Raw Escape Hatches

• figma_get_file
• figma_get_nodes

Use these only when raw Figma JSON is explicitly needed or when debugging the extension. They may return very large responses and reduce answer quality.

Output Limits

Processed tools enforce compact defaults and may include metadata.truncated: true plus nextSteps, for example:

• Call figma_get_node_summary with a deeper depth.
• Inspect a specific child node by ID.
• Enable includeComponentInternals=true for a focused component instance.

Notes

• Figma API is rate-limited; batch node IDs where supported.
• Large responses may be truncated; narrow to a specific child node when needed.
• This integration is read-only and cannot modify Figma files.
• Live Figma selection is not implemented. True Dev Mode-style live selection requires a future local Figma plugin/bridge.