首页/MCP Server 目录/arinspunk/claude-talk-to-figma-mcp

arinspunk/claude-talk-to-figma-mcp

❌ 实测失败serverTypeScriptMIT★ 676

A Model Context Protocol (MCP) that allows Claude Desktop and other AI tools (Claude Code, Cursor, Antigravity, etc.) to read, analyze, and modify Figma designs

实测结果

失败

沙箱内启动/握手失败，下方给出失败阶段与原始报错。失败阶段：initialize，错误类型：write_denied。这类信息对判断“是不是我这边的环境问题”很有用，原始报错见下。

error_type: write_denied

fatal: could not create work tree dir 'claude-talk-to-figma-mcp': Read-only file system

💥 Launcher error: Error: Command failed: git clone https://github.com/arinspunk/claude-talk-to-figma-mcp.git
    at genericNodeError (node:internal/errors:983:15)
    at wrappedFn (node:internal/errors:537:14)
    at checkExecSyncError (node:child_process:916:11)
    at execSync (node:child_process:988:15)
    at main (file:///work/node_modules/claude-talk-to-figma-mcp/scripts/launcher.js:95:9)
    at process.processTicksAndRejections (node:internal/process/task_queues:103:5) {
  status: 128,
  signal: null,
  output: [ null, null, null ],
  pid: 14,
  stdout: null,
  stderr: null
}

安装方式: npx
包名: claude-talk-to-figma-mcp
解析依据: npm registry match (has bin)
沙箱: docker
总耗时: 35769 ms
安装耗时: 34527 ms
工作目录: 28 MB
探针版本: 1.0.0
实测时间: 2026-11-11 15:30:46

如何使用

❌ 实测失败write_denied · fatal: could not create work tree dir 'claude-talk-to-figma-mcp': Read-only file system；下面配置仅供参考，跑不跑得通取决于你的环境

1安装 / 启动命令npm / npx

npx -y claude-talk-to-figma-mcp

2客户端配置Claude Desktop / Cursor

{
  "mcpServers": {
    "arinspunk-claude-talk-to-figma-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "claude-talk-to-figma-mcp"
      ]
    }
  }
}

整段粘进 claude_desktop_config.json 的 mcpServers 里（Cursor 用 ~/.cursor/mcp.json），保存后重启客户端。

安装方式来源：npm registry match (has bin)，沙箱实测未通过。

打开 GitHub 仓库 ↗MCP 体检工具

仓库信息

全名: arinspunk/claude-talk-to-figma-mcp
分类: server
Stars: 676
Forks: 134
主要语言: TypeScript
License: MIT
最近推送: 2026-04-18
仓库创建: 2025-04-04
首次收录: 2026-11-10
最近抓取: 2026-11-10

#antigravity#claude#claude-ai#claude-code#cursor#figma#githubcopilot#mcp#mcp-server#windsurf-ai