本文へ移動
cccskills
無料GitHub で公開

figma-design-analyzer

Extract and analyze Figma designs to create structured design specifications. Use this skill when you need to analyze Figma nodes, extract design properties (layout, spacing, typography), classify components using Atomic Design principles, and generate design specification documents that can be used for implementation.

インストール方法を見る

含まれるファイル(6)

  • SKILL.md15.6 KB
  • references/atomic-design-classification.md10.5 KB
  • references/layout-analysis-guide.md5.6 KB
  • references/semantic-html-guide.md14.4 KB
  • references/spacing-extraction-guide.md6.7 KB
  • references/typography-extraction-guide.md10.8 KB

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 Design Analyzer

Overview

Systematically analyze Figma designs and extract structured design specifications. This skill focuses on understanding design intent, extracting accurate properties, and creating comprehensive design specification documents that serve as blueprints for implementation.

Use this skill when:

  • Analyzing Figma designs to understand structure and properties
  • Extracting design tokens and component specifications
  • Classifying components according to Atomic Design principles
  • Creating design specification documents for developers
  • Comparing existing implementations with Figma designs

Core Workflow

Step 1: Extract Figma Design Information

When given a Figma URL or node ID, use Figma MCP tools to extract design data:

  1. Parse the Figma URL to extract node ID:

    URL: https://figma.com/design/fileKey/fileName?node-id=1-2
    Node ID: 1:2 (replace - with :)
    
  2. Extract node design context:

    mcp__figma-dev__get_design_context(
      nodeId: "1:2",
      dirForAssetWrites: "/absolute/path/to/project/public/assets",
      clientLanguages: "typescript",
      clientFrameworks: "react,nextjs"
    )
    

    This returns the design code, including HTML structure and CSS properties.

  3. Get component metadata for structure overview:

    mcp__figma-dev__get_metadata(
      nodeId: "1:2",
      clientLanguages: "typescript",
      clientFrameworks: "react,nextjs"
    )
    

    Use this to understand the component hierarchy before detailed analysis.

  4. Check for Figma variables (design tokens):

    mcp__figma-dev__get_variable_defs(
      nodeId: "1:2",
      clientLanguages: "typescript",
      clientFrameworks: "react,nextjs"
    )
    

    Extract design system tokens for colors, typography, spacing, etc.

  5. Verify Code Connect mappings:

    mcp__figma-dev__get_code_connect_map(
      nodeId: "1:2",
      clientLanguages: "typescript",
      clientFrameworks: "react,nextjs"
    )
    

    Check if the component is already mapped to existing code.

Step 2: Handle Component Instances and Variants

Important: When the node is a component instance:

  1. Identify if it's an instance:

    • Check metadata for "component instance" type
    • Note the main component ID
  2. Get the main component:

    • Use get_design_context on the main component node
    • Extract the base structure and properties
  3. Extract instance-specific properties:

    • Get the instance's specific size, device, and variant settings
    • Note any overrides (text, images, colors, etc.)
  4. Handle variants:

    • If the main component uses variants, identify the variant mode
    • Extract the correct variant values for the instance
    • Document variant properties for props mapping

Step 3: Classify the Component

Use Atomic Design principles to classify the component:

  1. Read the classification guide:

    Read references/atomic-design-classification.md
    
  2. Apply the decision tree:

    • Atom: Indivisible, single purpose (button, input, icon, text label)
    • Molecule: 2-5 atoms, simple combination (search bar, form field)
    • Organism: Complex, multiple molecules/atoms (header, card, form)
  3. Determine the component category:

    • Document: atomicCategory: Atom | Molecule | Organism

Step 4: Analyze Layout and Spacing

Extract accurate layout information:

  1. Read layout analysis guide:

    Read references/layout-analysis-guide.md
    
  2. Determine layout method:

    • Auto Layout → Flexbox
      • Check: direction (row/column), justify-content, align-items
    • Grid structure → CSS Grid
      • Check: grid-template-columns, grid-template-rows, gap
    • Overlapping elements → Absolute positioning
      • Check: position, top, left, z-index
  3. Extract spacing values:

    Read references/spacing-extraction-guide.md
    
    • Padding: Internal spacing (px, py, pt, pr, pb, pl)
    • Gap: Spacing between flex/grid items
    • Margin: External spacing (rarely used, prefer gap)
  4. Map to Tailwind classes:

    • Use standard Tailwind scale when possible (p-4, gap-6, etc.)
    • Use arbitrary values [value] for custom sizes (p-[18px])
  5. Document in specification:

    ## Layout
    - Method: Flexbox
    - Direction: row
    - Justify: center
    - Align: center
    
    ## Spacing
    - Padding: px-6 py-3
    - Gap: gap-2
    - Border Radius: rounded-lg
    

Step 5: Extract Typography

Extract accurate text styling:

  1. Read typography guide:

    Read references/typography-extraction-guide.md
    
  2. Extract font properties:

    • Font family (font-sans, font-serif, custom)
    • Font size (text-sm, text-base, text-lg, etc.)
    • Font weight (font-normal, font-semibold, font-bold)
    • Line height (leading-tight, leading-normal, etc.)
    • Letter spacing (tracking-tight, tracking-wide, etc.)
    • Text color (text-gray-900, text-white, etc.)
    • Text alignment (text-left, text-center, text-right)
    • Text decoration (underline, line-through)
    • Text transform (uppercase, lowercase, capitalize)
  3. Map to Tailwind classes:

    • Use standard Tailwind typography classes
    • Check for design tokens in Figma variables
    • Use arbitrary values for custom sizes
  4. Document in specification:

    ## Typography
    - Font: text-base font-semibold
    - Color: text-white
    - Line Height: leading-normal
    - Alignment: text-center
    

Step 6: Choose Semantic HTML

Select appropriate HTML elements:

  1. Read semantic HTML guide:

    Read references/semantic-html-guide.md
    
  2. Apply semantic rules:

    • Sections: <header>, <nav>, <main>, <aside>, <footer>
    • Headings: <h1> to <h6> (respect hierarchy)
    • Interactive: <button> for actions, <a> for navigation
    • Content: <article>, <section> for grouping
    • Forms: <form>, <input>, <label>, <select>, <textarea>
    • Lists: <ul>, <ol>, <li> for lists
    • Media: <img>, <video>, <audio>, <figure>, <figcaption>
  3. Consider accessibility:

    • Add ARIA attributes when semantic HTML is insufficient
    • Ensure keyboard accessibility
    • Verify color contrast ratios
  4. Document in specification:

    ## Semantic HTML
    - Root Element: button
    - ARIA: role="button" aria-label="Submit"
    

Step 7: Extract Colors and Visual Properties

Extract color and visual styling:

  1. Extract color properties:

    • Background color (bg-blue-500, bg-gradient-to-r, etc.)
    • Text color (text-white, text-gray-900, etc.)
    • Border color (border-gray-300, etc.)
  2. Extract visual properties:

    • Border radius (rounded, rounded-lg, rounded-full, etc.)
    • Border width (border, border-2, etc.)
    • Box shadow (shadow, shadow-md, shadow-lg, etc.)
    • Opacity (opacity-50, opacity-100, etc.)
  3. Check Figma variables:

    • Look for design token mappings
    • Use semantic color names when available
  4. Document in specification:

    ## Colors
    - Background: bg-blue-500
    - Hover: hover:bg-blue-600
    - Text: text-white
    - Border: border-transparent
    
    ## Visual Properties
    - Border Radius: rounded-lg
    - Shadow: shadow-md
    

Step 8: Handle Responsive Designs

When multiple Figma nodes represent different device sizes:

  1. Extract all node IDs from provided URLs

    • Desktop node
    • Tablet node
    • Mobile node
  2. Analyze each node:

    • Layout changes (grid columns, flex direction)
    • Spacing changes
    • Typography changes (font size, line height)
    • Hidden/shown elements
  3. Document responsive variations:

    ## Responsive Behavior
    - Mobile (default): flex-col gap-4 text-sm
    - Tablet (md): md:flex-row md:gap-6 md:text-base
    - Desktop (lg): lg:gap-8 lg:text-lg
    

Step 9: Generate Design Specification Document

Create a comprehensive design specification:

  1. Use Markdown format with front matter:

    ---
    componentName: Button
    atomicCategory: Atom
    semanticElement: button
    figmaNodeId: "123:456"
    figmaUrl: "https://figma.com/design/..."
    ---
    
    # Button Component Specification
    
    ## Overview
    A primary action button component.
    
    ## Atomic Design Classification
    - Category: Atom
    - Reason: Single, indivisible interactive element
    
    ## Layout
    - Method: Flexbox
    - Direction: row
    - Justify: center
    - Align: center
    
    ## Spacing
    - Padding: px-6 py-3
    - Gap: gap-2
    - Border Radius: rounded-lg
    
    ## Typography
    - Font: text-base font-semibold
    - Color: text-white
    - Line Height: leading-normal
    - Alignment: text-center
    
    ## Colors
    - Background: bg-blue-500
    - Hover: hover:bg-blue-600
    - Active: active:bg-blue-700
    - Disabled: disabled:bg-gray-300
    
    ## Visual Properties
    - Border: border-transparent
    - Shadow: shadow-md
    - Transition: transition-colors duration-200
    
    ## Semantic HTML
    - Root Element: button
    - ARIA: aria-label when no text content
    
    ## Variants
    - variant: primary | secondary | tertiary
      - primary: bg-blue-500 text-white
      - secondary: bg-white text-blue-500 border-blue-500
      - tertiary: bg-transparent text-blue-500
    - size: small | medium | large
      - small: px-4 py-2 text-sm
      - medium: px-6 py-3 text-base
      - large: px-8 py-4 text-lg
    
    ## States
    - Default: As specified above
    - Hover: Darker background
    - Active: Even darker background
    - Disabled: Gray background, reduced opacity
    - Focus: Ring outline for keyboard navigation
    
    ## Responsive Behavior
    - Mobile: Full width, smaller padding
    - Tablet: Inline, medium padding
    - Desktop: Inline, larger padding
    
    ## Accessibility
    - Keyboard: Focusable, Enter/Space to activate
    - Screen Reader: Announced as "button"
    - Color Contrast: WCAG AA compliant
    
    ## Assets
    - Icons: /public/assets/icon-arrow.svg (if any)
    - Images: None
    
    ## Notes
    - Always use semantic <button> element
    - Avoid using <a> styled as button
    - Include loading state if used for async actions
    
  2. Save the specification:

    • Filename: {ComponentName}.design.md
    • Location: /design-specs/ or project-specific location

Special Workflows

Comparing Design with Implementation

When asked to compare an existing component with Figma:

  1. Read the existing component code
  2. Extract Figma design properties using MCP tools
  3. Compare systematically:
    • Layout method (Flexbox, Grid, etc.)
    • Spacing values (padding, gap, margin)
    • Typography (font size, weight, line height, color)
    • Colors and backgrounds
    • Border radius, shadows, etc.
  4. Generate a diff report:
    ## Design Differences Report
    
    ### Component: Button
    
    #### Spacing
    - Design: px-6 py-3 gap-2
    - Implementation: px-4 py-2 gap-3
    - Status: ❌ Mismatch
    - Impact: Button appears smaller and less comfortable
    
    #### Typography
    - Design: text-base font-semibold
    - Implementation: text-base font-semibold
    - Status: ✅ Match
    
    #### Colors
    - Design: bg-blue-500 hover:bg-blue-600
    - Implementation: bg-blue-600 hover:bg-blue-700
    - Status: ❌ Mismatch
    - Impact: Button is darker than intended
    
    ### Recommendations
    1. Update padding to px-6 py-3
    2. Update gap to gap-2
    3. Update background colors to match design
    
  5. Document required changes in the specification

Breaking Down a Full Page

When given a full page Figma node:

  1. Get page metadata to understand structure

  2. Identify major sections:

    • Header
    • Hero
    • Content sections
    • Footer
  3. For each section:

    • Classify as Organism
    • Identify child Molecules and Atoms
    • Extract design properties
    • Create individual specifications
  4. Create component tree:

    Page (Template)
    ├─ Header (Organism)
    │  ├─ Logo (Atom)
    │  ├─ Navigation (Molecule)
    │  └─ UserMenu (Molecule)
    ├─ Hero (Organism)
    │  ├─ Heading (Atom)
    │  ├─ Paragraph (Atom)
    │  └─ CTAButton (Atom)
    ├─ Features (Organism)
    │  └─ FeatureCard (Molecule) × 3
    │     ├─ Icon (Atom)
    │     ├─ Title (Atom)
    │     └─ Description (Atom)
    └─ Footer (Organism)
       ├─ Logo (Atom)
       └─ Links (Molecule)
    
  5. Generate specifications from bottom-up:

    • Create Atom specifications first
    • Build Molecule specifications
    • Assemble Organism specifications
    • Compose Page specification with assembly instructions

Best Practices

  1. Always read guides before starting:

    • Layout analysis guide for layout decisions
    • Spacing guide for accurate spacing extraction
    • Typography guide for text styling
    • Atomic Design guide for classification
    • Semantic HTML guide for element selection
  2. Prioritize pixel-perfect accuracy:

    • Use exact spacing values from Figma
    • Match typography precisely
    • Verify colors match design tokens
  3. Document thoroughly:

    • Include all variants and states
    • Document responsive behavior
    • Note accessibility considerations
    • Include designer notes if available
  4. Consider implementation:

    • Map to Tailwind classes for easy implementation
    • Document component hierarchy
    • Note any complex interactions
  5. Maintain consistency:

    • Use consistent naming conventions
    • Follow project-specific patterns
    • Align with design system when available

Resources

This skill includes comprehensive guides:

references/

  • layout-analysis-guide.md: Criteria for choosing Flexbox, Grid, absolute positioning
  • spacing-extraction-guide.md: Methods for extracting padding, gap, margin values
  • typography-extraction-guide.md: Guide for extracting font properties
  • atomic-design-classification.md: Classification criteria for Atoms, Molecules, Organisms
  • semantic-html-guide.md: Recommendations for choosing appropriate HTML elements

Usage: Always read relevant guides before starting analysis to ensure accuracy.

Common Pitfalls

  1. Not handling component instances properly:

    • Always get the main component when working with instances
    • Extract instance-specific properties separately
  2. Ignoring variants:

    • Check for variant properties in Figma
    • Document all variant options
  3. Using wrong layout method:

    • Read the layout guide
    • Don't default to Flexbox for everything
  4. Skipping semantic HTML:

    • Don't use <div> for everything
    • Consider accessibility and SEO
  5. Forgetting responsive behavior:

    • Check for responsive variants in Figma
    • Document breakpoint changes
  6. Not checking existing code:

    • Always check Code Connect mappings
    • Avoid duplicating existing specifications

Output Format

The final output should be a Markdown document with:

  1. Front matter with metadata
  2. Overview section
  3. Design properties organized by category
  4. Variants and states documentation
  5. Responsive behavior notes
  6. Accessibility considerations
  7. Assets references
  8. Implementation notes

This specification document will be used by the nextjs-react-implementation skill to generate actual code.


This skill systematically analyzes Figma designs and creates structured specifications for accurate implementation.

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

android-api-client

無料日本語概要

Retrofit + OkHttpを使用したAPIクライアント実装のベストプラクティスを提供する。 「API実装」「APIクライアント」「Retrofit」「OkHttp」「ネットワーク」と言われたら使う。

xtone/ai_development_tools32026年10月7日 更新

android-architecture

無料日本語概要

Androidプロジェクトのアーキテクチャ設計を支援。プロジェクト特性に応じてMVI/MVVM/シンプルComposeから最適なパターンを提案し、パッケージ構造の設計と既存プロジェクトの分析を行う。

xtone/ai_development_tools32026年10月7日 更新

android-crash-analyzer

無料日本語概要

Firebase Crashlyticsのスタックトレースを解析し、トリアージ分類→根本原因分析→修正提案を出力するスキル。/crash-analyze コマンド、クラッシュ解析、スタックトレース分析、ANR分析、Crashlytics、クラッシュレポート、またはFirebase crashに関連するリクエストで使用される。クラッシュ対応のワークフローを標準化し、対応品質と速度を向上させる。ANR/Non-fatal/Fatal全タイプ対応。

xtone/ai_development_tools32026年10月7日 更新

Android data layer implementation guide. Provides comprehensive patterns for Room Database (Entity, DAO, Migration), DataStore (Preferences), Repository layer design, and common pitfalls. Use when implementing local data storage, cache strategies, or data source integration in Android projects.

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。

xtone/ai_development_tools32026年10月7日 更新

android-test-runner

無料日本語概要

重要: ユーザーがAndroidテスト実行をリクエストした場合、常にこのスキルを最初に使用してください。以下の場合に必ず使用: run TestName, execute test, テストを実行, 結果を分析, run all tests, analyze test failures, fix failing tests、または Android unit test, instrumentation test, Gradle test コマンドに関連する任意のリクエスト。./gradlew test や Bash コマンドを直接使用しないでください - 常にこのスキルに委譲してください。Multi-variantプロジェクト、JAVA_HOME セットアップ、一般的なテストパターンに対応しています。

xtone/ai_development_tools32026年10月7日 更新

android-ui-guidelines

無料日本語概要

Jetpack ComposeでUIを実装する時に参照するコーディングガイドライン。 「Android UI」「Compose」「Jetpack」「画面実装」「UIコンポーネント」「レイアウト」と言われたら使う。 Android/Kotlin UI実装のベストプラクティス、アンチパターン、実装パターンを提供する。

xtone/ai_development_tools32026年10月7日 更新

xtone のスキルをすべて見る

このスキルの問題を報告する