convez.ai
MCP ServerDesign & UIIntermediário
hsiangyilu

Design System Builder

hsiangyilu/design-system-builder

Construa sistemas de design do zero com IA, do brand até Figma

Ver no GitHub
InicialPluginScore · de 0 a 100

MCP server que guia a construção de sistemas de design completos através de um workflow estruturado de 6 fases. Integra Supa Palette para cores, Typescale para tipografia e Figma MCP para variáveis, gerando tokens.json, configuração Tailwind e documentação automaticamente.

Instalação

$npx convez-ai hsiangyilu/design-system-builder

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
0
Acompanhando
0
Saúde
28%
Último commit
há 4 meses
Versão
—
Issues abertas
0
Contribuidores
2

Casos de uso

  • Gerar paletas de cores com análise de contraste automática
  • Definir escalas tipográficas e recomendações de pareamento de fontes
  • Criar tokens.json e config Tailwind prontos para projetos Next.js
  • Publicar variáveis primitivas e semânticas no Figma via MCP
  • Documentar decisões de design com rationale estruturado

Veredito

Ferramenta sólida para times que precisam padronizar identidade visual em código e Figma simultaneamente. Requer Figma MCP ativo e projeto Next.js/Tailwind, mas entrega documentação técnica completa e reutilizável.

README

A tradução para o português deste README ainda está na fila.

🎨 Design System Builder Skill

A Claude skill that guides you through building a complete design system from scratch — step by step, tool by tool.

繁體中文版說明見下方 ↓


What it does

When you say "build a design system" or "set up design tokens", this skill activates and walks you through a structured 6-phase workflow:

PhaseWhat happensWho does it
1. Brand decisionsLock in personality, color direction, usage contextYou
2. Color systemGenerate full color scale via Supa Palette, review contrastAI + You
3. TypographySet type scale via Typescale, pick font pairingAI + You
4. Token integrationOutput tokens.json + Tailwind config automaticallyAI
5. Figma VariablesPush Primitive + Semantic layers into Figma via MCPAI
6. DocumentationGenerate design-system-doc.md with Usage / Don't / RationaleAI + You

Tools used

  • Supa Palette — color scale generation
  • Typescale — type scale generation
  • Figma MCP — Variables + visual showcase pages
  • Next.js / Tailwind CSS — frontend token integration

Outputs

For every project, the skill produces:

tokens.json                 ← design source of truth (primitive + semantic layers)
tailwind.config.ts          ← drop into your Next.js project
design-system-doc.md        ← Usage / Don't / Design Rationale
Figma Variables             ← 🎨 Primitive + 🖌 Semantic + 📐 Typography collections
Figma showcase pages        ← color swatches + typography scale + component preview

Install

Download design-system-builder.skill and place it in:

# Global (available in all projects)
~/.claude/skills/

# Per-project
.agents/skills/

Then in Claude Code or Claude chat, just say:

"幫我建立設計系統" / "build a design system" / "I need design tokens"


Example output

See examples/y2k-snap/ for a real project walkthrough — a Y2K aesthetic photo editing app built with this skill.

Includes:

  • tokens.json — full primitive + semantic color and typography tokens
  • tailwind.config.ts — ready to use
  • screenshots/ — Figma Variables panel + visual showcase pages

Trigger phrases

The skill activates on:

  • 製作設計系統 / 建立 design system
  • 我需要 design token / brand token
  • 幫我定義色彩系統 / 建立字型規範
  • 設計系統從零開始 / token 整合
  • build a design system / set up design tokens
  • my UI looks inconsistent / I need a color system

Requires

  • Claude with Figma MCP connected (for Phases 5)
  • Figma Desktop App with fonts installed locally
  • Next.js + Tailwind CSS project (for token integration)


繁體中文說明

這個 Skill 做什麼

說出「製作設計系統」或「我需要 design token」,skill 自動啟動,帶你走完六個 Phase:

  1. 品牌決策 — 確認個性關鍵字、色彩方向、使用情境(你決定)
  2. 色彩系統 — 用 Supa Palette 生成完整 scale,AI 審查 contrast(你 + AI)
  3. 字型系統 — 用 Typescale 設定階層,AI 推薦字型配對(你 + AI)
  4. Token 整合 — 自動產出 tokens.json + Tailwind config(AI 全自動)
  5. Figma Variables — 透過 Figma MCP 建立 Primitive + Semantic 兩層(AI 全自動)
  6. 文件化 — 產出含 Usage / Don't / Design Rationale 的設計文件(AI 起草,你補 Rationale)

安裝方式

下載 design-system-builder.skill,放到:

# 全局安裝(所有專案都能用)
~/.claude/skills/

# 單一專案安裝
.agents/skills/

產出物

每個專案結束後你會拿到:

  • tokens.json — primitive + semantic 兩層 token
  • tailwind.config.ts — 直接貼進 Next.js 專案
  • design-system-doc.md — 完整設計文件含決策原因
  • Figma Variables — 三個 collection 全部建好
  • Figma 視覺展示頁 — 色票 + 字型 scale + 元件預覽

注意事項

  • Phase 5 需要 Figma MCP 連線
  • 字型需本地安裝才能在 Figma 正確顯示
  • Design Rationale 只有你能寫,AI 不知道你的決策脈絡

License

MIT — fork it, adapt it, make it yours.

Ver o original no GitHub

Bom para

Relacionados