本文へ移動
cccskills
無料GitHub で公開

figma

Figma design-to-code workflow. Read Figma files, inspect components, extract design tokens, and translate designs to implementation. Use when working with Figma designs, extracting styles, or building design systems.

インストール方法を見る

含まれるファイル(10)

  • SKILL.md7.9 KB
  • commands/figma.md105 B
  • hooks/post-execute.cjs163 B
  • hooks/pre-execute.cjs269 B
  • references/research-requirements.md614 B
  • rules/figma.md749 B
  • schemas/input.schema.json578 B
  • schemas/output.schema.json381 B
  • scripts/main.cjs593 B
  • templates/implementation-template.md272 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Figma

Extract design information from Figma files for implementation. Covers design token extraction, component inspection, and design-to-code translation.

When to Use

  • Translating Figma designs into frontend code
  • Extracting design tokens (colors, typography, spacing) from Figma
  • Inspecting Figma component structure for implementation planning
  • Building or updating a design system from Figma sources
  • Mapping Figma auto-layout properties to CSS

Prerequisites

  • Figma Personal Access Token (set as FIGMA_ACCESS_TOKEN env var)
  • Figma file URL or file key

Core API Endpoints

Get File

curl -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/files/FILE_KEY"

Get File Nodes (specific components)

curl -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/files/FILE_KEY/nodes?ids=NODE_ID1,NODE_ID2"

Get Image Exports

curl -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/images/FILE_KEY?ids=NODE_ID&format=svg"

Get File Styles

curl -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/files/FILE_KEY/styles"

Get File Components

curl -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/files/FILE_KEY/components"

Design Token Extraction Workflow

Step 1: Fetch File Styles

# Get all published styles
curl -s -H "X-Figma-Token: $FIGMA_ACCESS_TOKEN" \
  "https://api.figma.com/v1/files/FILE_KEY/styles" | jq '.meta.styles'

Step 2: Map to CSS Variables

/* Colors */
:root {
  --color-primary: #2563eb; /* From Figma: Primary/500 */
  --color-secondary: #7c3aed; /* From Figma: Secondary/500 */
  --color-surface: #ffffff; /* From Figma: Surface/Default */
}

/* Typography */
:root {
  --font-heading: 'Inter', sans-serif;
  --font-body: 'Inter', sans-serif;
  --text-xl: 1.25rem; /* From Figma: Heading/XL */
  --text-base: 1rem; /* From Figma: Body/Base */
  --leading-tight: 1.25;
  --leading-normal: 1.5;
}

/* Spacing */
:root {
  --space-1: 0.25rem; /* 4px */
  --space-2: 0.5rem; /* 8px */
  --space-4: 1rem; /* 16px */
  --space-8: 2rem; /* 32px */
}

Step 3: Auto-Layout to CSS Mapping

Figma Auto-LayoutCSS Equivalent
Horizontaldisplay: flex; flex-direction: row;
Verticaldisplay: flex; flex-direction: column;
Space betweenjustify-content: space-between;
Packed (start)justify-content: flex-start;
Gap: 16gap: 1rem;
Padding: 16,24padding: 1rem 1.5rem;
Fill containerflex: 1; or width: 100%;
Hug contentswidth: fit-content;

Step 4: Component Inspection

For each Figma component, extract:

  1. Name - Component name and variant properties
  2. Structure - Child node tree (frames, text, images)
  3. Styles - Applied color, text, and effect styles
  4. Constraints - Responsive behavior rules
  5. Variants - All component variants and their properties

Iron Laws

  1. ALWAYS extract design tokens before building components — building without tokens leads to hardcoded values that break when the design system updates.
  2. NEVER hardcode Figma colors/spacing — always map to CSS variables or design token constants that reference the Figma source.
  3. ALWAYS map Figma auto-layout to CSS flexbox — auto-layout is a direct 1:1 mapping to flexbox; using other layout methods loses design intent.
  4. NEVER export raster images when SVG is available — SVGs scale cleanly and reduce bundle size; only use PNG/JPEG for photographs.
  5. ALWAYS preserve Figma component naming in code component names — divergent naming breaks the design-development feedback loop.

Anti-Patterns

Anti-PatternWhy It FailsCorrect Approach
Hardcoded hex colorsBreaks on design system updateUse CSS variables mapped from Figma tokens
Ignoring auto-layoutLayout doesn't match designMap auto-layout 1:1 to flexbox/grid
Pixel-perfect obsessionWastes time, ignores responsiveMatch design intent, not exact pixels
Skipping component variantsMissing states (hover, disabled)Extract ALL Figma variants
Manual token extractionError-prone, doesn't scaleUse Figma REST API for programmatic extraction

Figma MCP Server Tools (figma/mcp-server-guide)

When the official Figma MCP server is running, these 12 tools are available:

ToolPurpose
get_design_contextGet full design context for a file/node (components, styles, layout)
generate_figma_designGenerate new designs from natural language description
get_variable_defsExtract all design tokens (colors, spacing, typography variables)
get_screenshotCapture rendered screenshot of a frame/component
code_connect_mapMap Figma components to their code counterparts
create_design_system_rulesAuto-generate design system documentation
get_figjamAccess FigJam board content and sticky notes
generate_diagramGenerate Figma diagram from Mermaid input
get_componentsList all published components in a file
get_stylesExtract all defined styles (color, text, effect)
export_assetsExport assets as PNG/SVG/PDF with scale options
get_annotationsRetrieve design annotations and spec notes
get_metadataGet XML layer representation of a file/node for programmatic parsing
whoamiGet current user identity, plan tier, and available seat count

MCP Setup: Install Figma Desktop app -> Settings -> Enable MCP server Auth: Uses active Figma session (no API key needed for local MCP) Key workflow: get_design_context -> code_connect_map -> export_assets for design-to-code Identity check: Use whoami to verify plan tier before calling enterprise-only tools

Memory Protocol (MANDATORY)

Before starting: Read .claude/context/memory/learnings.md

After completing:

  • New pattern -> .claude/context/memory/learnings.md
  • Issue found -> .claude/context/memory/issues.md
  • Decision made -> .claude/context/memory/decisions.md

ASSUME INTERRUPTION: If it's not in memory, it didn't happen.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Use when you want to improve response quality through meta-cognitive reasoning. Applies 15+ reasoning methods to reconsider and refine initial outputs.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

N-round opposing-stance debates for trade-off analysis. Assigns pro/con roles to agents, runs structured debate rounds with quality scoring, and produces a moderator synthesis with confidence-rated recommendation. Generalizable to architecture, technology, security, and design decisions.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Force adversarial code review stance that eliminates confirmation bias — reviewer must find issues or re-analyze

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Creates specialized AI agents on-demand when no existing agent matches a request. Use when the Router cannot find a suitable agent for a task. Enables self-evolution by generating persistent agents.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

LLM-as-judge evaluation framework with 5-dimension rubric (accuracy, groundedness, coherence, completeness, helpfulness) for scoring AI-generated content quality with weighted composite scores and evidence citations

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

oimiragieo のスキルをすべて見る

このスキルの問題を報告する