JS Development MCP Server
A Model Context Protocol server for JavaScript/TypeScript development, providing tools to run scripts, tests, builds, and dependency management with framework and package manager detection.
README
JS Development MCP Server
A Model Context Protocol (MCP) server for better JS/TS development. It enables
- accurate scripts use
- project tooling discovery
- improved background tasks
Better context leads to less friction and struggle with project tooling.
<img src="https://github.com/Tymek/js-mcp/raw/master/js-mcp-demo.gif" alt="JS MCP Demo" style="max-width: 500px;">
Warning: Using this has a YOLO vibe coding factor of arbitrary 8/10 – No warranty if this breaks your project, computer, or your life, but it probably won't. Don't leave your agents unsupervised. This server allows you to run npm scripts in VSCode copilot without manual terminal command approval every time, so be careful.
Quick Start
npx js-mcp
For VSCode / Claude Desktop:
{
"mcpServers": {
"js-mcp": {
"command": "npx",
"args": ["js-mcp"]
}
}
}
Instructions for LLM Agents & AI Assistants
🤖 Agent Instructions for JS/TS Projects
In Javascript or Typescript projects, PREFER TO USE TOOLS INSTEAD OF TERMINAL COMMANDS:
npm test→ run-tests (framework detection, better errors)npm run build→ build-project (optimization detection)npm install→ install-deps (auto-detects npm/yarn/pnpm)npm run <script>→ run-script (background detection, args support)- Checking package.json → list-scripts or project-status
Benefits: No terminal approval prompts, enhanced error context, framework awareness, proper background task management.
Workflow: Start with
get-helpfor guidance, useproject-statusfor unknown projects, then use specific tools.
Tools
- get-help - Get intelligent recommendations for which tools to use
- run-tests - Run project tests with framework detection and better errors
- build-project - Build projects with optimization detection and framework awareness
- install-deps - Install dependencies with auto-detected package manager (npm/yarn/pnpm)
- run-script - Run npm scripts with background detection and argument support
- list-scripts - List and categorize all available npm scripts
- project-status - Check comprehensive project health and framework detection
Key Features
- Framework Detection: Next.js, React, Vue.js, Express, NestJS, and more
- Background Task Management: Auto-detection for dev servers, manual control, output capture
- Smart Error Handling: Contextual suggestions and dependency checking
- Package Manager Detection: Auto-detects npm, yarn, or pnpm from lock files
- Custom Timeout Support: Configure timeouts (1s-5m) for long-running commands, with smart defaults
Background Task Auto-Detection
Always Background: Scripts containing dev, start, serve, watch
Never Background: Scripts starting with cleanup, clean, reset, install, build, compile, bundle, lint, format, test, deploy, publish, release, prepare, verify, check, validate, audit, update, upgrade, migration, migrate, seed, init, setup, config
Development
Local development:
git clone <repo>
npm install
npm run build
npm start
License
GPL-3.0 License
Contributing
Go ahead. Feel free to extend this project with additional tools as needed. I can't promise to review or merge PRs. This codebase is a gift, not a commitment.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。