R
rafaelmatosdasilva/rms-design-system-engine
Maintains your design system across design, code and AI
在 GitHub 查看 ↗链接已于 2026-11-12 核查 · HTTP 200 · 打开的是厂商自己的官网
官方说明Maintains your design system across design, code and AI. Generates contracts, documentation and style guides from real sources, checks parity, component usage and accessibility against explicit rules, and keeps products aligned as tokens, components and requirements evolve. Works inside Claude Code.
- ★ 数取值
- 1 ★ · rafaelmatosdasilva/rms-design-system-engine · 取自 GitHub 搜索 API · （2026-11-13 20:00:50 核对）
- 来源明细
- 从该仓库的 GitHub 页面直接读取 · 条目类型：开源项目
- 收录与活跃
- 收录于 2026-10-29 · 最近见到 2026-11-14 · 共见到 17 次 · 被 2 个来源列出
- 评分构成
- 综合分 203 · 1 ★ · 2 个来源 · repo
常见问题（答案都来自我们的记录）
它适合谁 / 用来做什么？
我们把它归在这些标签下: 开源项目。标签来自条目自身，不是我们强加的分类。
谁维护它、还在更新吗？
仓库
rafaelmatosdasilva/rms-design-system-engine；我们收录于 2026-10-29；最近见到 2026-11-14。这里没有"官方认证"或"质量评级"——我们只记录来源与时间。# The maintenance engine for AI-ready design systems
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
AI agents can build interfaces in minutes. The hard part is making sure what they build follows your design system, and keeps following it as tokens, components and requirements change.
RMS Design System Engine reads what you have (a Figma file, a codebase, or both) and turns it into documentation and component contracts that people and agents can use. It then checks every implementation against explicit rules, many of them measured in a real browser, instead of trusting the agent to get it right. Every check builds on the same shared knowledge, so what one task learns, the next one uses.
## Who it's for
Design-led teams working in Figma who want to turn their design system into working code, build prototypes with real components and verify accessibility.
Design and engineering teams who need to maintain parity between Figma and code, verify implementations against design intent and catch inconsistencies as designs and code evolve.
Code-first teams who want to improve the reliability, accessibility and maintainability of their codebase, document components, track design system health and ensure AI-generated code follows their implementation rules.
## What it does
Understand: Turn real design and code sources into structured design intent, component contracts, documentation and style guides that people and AI agents can use.
Verify: Run deterministic checks for design-code parity, component usage, accessibility and implementation rules, with actionable findings rather than relying on AI judgement alone.
Maintain: Detect inconsistencies and regressions as tokens, components, requirements and implementations change, keeping products aligned with the design system over time.
## Get started
Requirements: Claude Code, Node.js 22+ and Git.
Recommended: Google Chrome, so everything can be measured as a browser draws it, and TypeScript 5 (npm i -g typescript@5), so your components' JSX and props are read. Without them those checks are skipped; the installer tells you which are missing.
Ask Claude Code to run this command to install the engine:
bash
curl -fsSL https://raw.githubusercontent.com/rafaelmatosdasilva/rms-design-system-engine/main/install.sh | bash
Restart Claude Code and run:
text
/rms-design-system-engine set up this project
Once configured, ask Claude Code to check parity, audit accessibility, generate a style guide, build prototypes or investigate component usage.
RMS Design System Engine reads what you have (a Figma file, a codebase, or both) and turns it into documentation and component contracts that people and agents can use. It then checks every implementation against explicit rules, many of them measured in a real browser, instead of trusting the agent to get it right. Every check builds on the same shared knowledge, so what one task learns, the next one uses.
## Who it's for
Design-led teams working in Figma who want to turn their design system into working code, build prototypes with real components and verify accessibility.
Design and engineering teams who need to maintain parity between Figma and code, verify implementations against design intent and catch inconsistencies as designs and code evolve.
Code-first teams who want to improve the reliability, accessibility and maintainability of their codebase, document components, track design system health and ensure AI-generated code follows their implementation rules.
## What it does
Understand: Turn real design and code sources into structured design intent, component contracts, documentation and style guides that people and AI agents can use.
Verify: Run deterministic checks for design-code parity, component usage, accessibility and implementation rules, with actionable findings rather than relying on AI judgement alone.
Maintain: Detect inconsistencies and regressions as tokens, components, requirements and implementations change, keeping products aligned with the design system over time.
## Get started
Requirements: Claude Code, Node.js 22+ and Git.
Recommended: Google Chrome, so everything can be measured as a browser draws it, and TypeScript 5 (npm i -g typescript@5), so your components' JSX and props are read. Without them those checks are skipped; the installer tells you which are missing.
Ask Claude Code to run this command to install the engine:
bash
curl -fsSL https://raw.githubusercontent.com/rafaelmatosdasilva/rms-design-system-engine/main/install.sh | bash
Restart Claude Code and run:
text
/rms-design-system-engine set up this project
Once configured, ask Claude Code to check parity, audit accessibility, generate a style guide, build prototypes or investigate component usage.
收录徽章（可选）
想把收录标记放进 README 就复制这行，可选，不影响收录。
[](https://anclum.com/zh/s/rafaelmatosdasilva-rms-design-system-engine/)
- 来源
- github · https://github.com/rafaelmatosdasilva/rms-design-system-engine
- 按领域浏览
- 开源项目
- Traction
- 1 ★
- 收录来源
- 2 个独立来源
- 收录于
- 今天
同类推荐
所属领域与标签
每个标签都能点进去看同类的全部条目。数字是当前库里的真实条数。