Agent skill

design-to-code

Pixel-perfect Figma to React conversion using coderio. Generates production-ready code (TypeScript, Vite, TailwindCSS V4) with high visual fidelity. Features robust error handling, checkpoint recovery, and streamlined execution via helper script.

View SKILL.md on GitHub Repository
Stars 23,776
Forks 2,298

Install this agent skill to your Project

npx add-skill https://github.com/davila7/claude-code-templates/tree/main/cli-tool/components/skills/design-to-code

SKILL.md

Design to Code

High-fidelity UI restoration from Figma designs to production-ready React + TypeScript components. This SKILL uses a robust helper script to minimize manual errors and ensure pixel-perfect results.

Prerequisites

  1. Figma API Token: Get from Figma → Settings → Personal Access Tokens
  2. Node.js: Version 18+
  3. coderio: Installed in scripts/ folder (handled by Setup phase)

Workflow Overview

Phase 0: SETUP    → Create helper script and script environment
Phase 1: PROTOCOL → Generate design protocol (Structure & Props)
Phase 2: CODE     → Generate components and assets

Phase 0: Setup

Step 0.1: Initialize Helper Script

User Action: Run these commands to create the execution helper and isolate its dependencies.

bash
mkdir -p scripts

# 1. Copy script files
# Note: Ensure you have the 'skills/design-to-code/scripts' directory available
cp skills/design-to-code/scripts/package.json scripts/package.json
cp skills/design-to-code/scripts/coderio-skill.mjs scripts/coderio-skill.mjs

# 2. Install coderio in scripts directory (adjust version if needed)
cd scripts && pnpm install && cd ..

Step 0.2: Scaffold Project (Optional)

If starting a new project:

  1. Run: node scripts/coderio-skill.mjs scaffold-prompt "MyApp"
  2. AI Task: Follow the instructions output by the command to create files.

Phase 1: Protocol Generation

Step 1.1: Fetch Data

bash
# Replace with your URL and Token
node scripts/coderio-skill.mjs fetch-figma "https://figma.com/file/..." "figd_..."

Verify: process/thumbnail.png should exist.

Step 1.2: Generate Structure

  1. Generate Prompt:

    bash
    node scripts/coderio-skill.mjs structure-prompt > scripts/structure-prompt.md

  2. AI Task (Structure):

    • ATTACH: process/thumbnail.png (MANDATORY)
    • READ: scripts/structure-prompt.md
    • INSTRUCTION: "Generate the component structure JSON based on the prompt and the attached thumbnail. Focus on visual grouping. Use text content to name components accurately (e.g. 'SafeProducts', not 'FAQ')."
    • SAVE: Paste the JSON result into scripts/structure-output.json.

  3. Process Result:

    bash
    node scripts/coderio-skill.mjs save-structure

Step 1.3: Extract Props (Iterative)

  1. List Components:

    bash
    node scripts/coderio-skill.mjs list-components

  2. For EACH component in the list:

    a. Generate Prompt:

    bash
    node scripts/coderio-skill.mjs props-prompt "ComponentName" > scripts/current-props-prompt.md

    b. AI Task (Props):

    • ATTACH: process/thumbnail.png (MANDATORY)
    • READ: scripts/current-props-prompt.md
    • INSTRUCTION: "Extract props and state data. Be pixel-perfect with text and image paths."
    • SAVE: Paste the JSON result into scripts/ComponentName-props.json.

    c. Save & Validate:

    bash
    node scripts/coderio-skill.mjs save-props "ComponentName"
# If this fails, re-do step 'b' with better attention to the thumbnail

Phase 2: Code Generation

Step 2.1: Plan Tasks

bash
node scripts/coderio-skill.mjs list-gen-tasks

This outputs a list of tasks with indices (0, 1, 2...).

Step 2.2: Generate Components (Iterative)

For EACH task index (starting from 0):

  1. Generate Prompt:

    bash
    node scripts/coderio-skill.mjs code-prompt 0 > scripts/code-prompt.md
# Replace '0' with current task index

  2. AI Task (Code):

    • ATTACH: process/thumbnail.png (MANDATORY)
    • READ: scripts/code-prompt.md
    • INSTRUCTION: "Generate the React component code. Match the thumbnail EXACTLY. Use STRICT text content from input data, do not hallucinate."
    • SAVE: Paste the code block into scripts/code-output.txt.

  3. Save Code:

    bash
    node scripts/coderio-skill.mjs save-code 0
# Replace '0' with current task index

Step 2.3: Final Integration

Inject the root component into App.tsx. Use the path found in the last task of Phase 2.1.

Troubleshooting

  • "Props validation failed": The AI generated empty props. Check if process/thumbnail.png was attached and visible to the AI. Retry the props generation step.
  • "Module not found": Ensure node scripts/coderio-skill.mjs save-code was run for the child component before the parent component. Phase 2 must be done in order (0, 1, 2...).
  • "Visuals don't match": Did you attach the thumbnail? The AI relies on it for spacing and layout nuances not present in the raw data.

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

davila7/claude-code-templates

verl-rl-training

Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.

23,776 2,298
Explore
davila7/claude-code-templates

openrlhf-training

High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.

23,776 2,298
Explore
davila7/claude-code-templates

gguf-quantization

GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.

23,776 2,298
Explore
davila7/claude-code-templates

Claude Code Guide

Master guide for using Claude Code effectively. Includes configuration templates, prompting strategies "Thinking" keywords, debugging techniques, and best practices for interacting with the agent.

23,776 2,298
Explore
davila7/claude-code-templates

qdrant-vector-search

High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.

23,776 2,298
Explore
davila7/claude-code-templates

behavioral-modes

AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.

23,776 2,298
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results