Simon MCP Server

A Model Context Protocol (MCP) server that provides claude-like command execution capabilities for development tasks.

Features

• Natural Language Commands: Execute development tasks using natural language descriptions
• Project-Aware: Configurable project path restrictions for security
• Multiple Actions: Support for refactoring, development, analysis, debugging, and more
• Security: Built-in sandboxing and file access controls
• Easy Integration: Simple npm package installation and MCP configuration

Installation

Option 1: Install from npm (Recommended)

# Install globally
npm install -g simon-ai

# Or use with npx (no installation required)
npx simon-ai

Option 2: Build from source

1. Clone this repository:

git clone <repository-url>
cd simon-mcp

2. Install dependencies:

npm install

3. Build the project:

npm run build

MCP Configuration

Add to your MCP client configuration:

{
  "mcpServers": {
    "simon": {
      "command": "npx",
      "args": ["-y", "@simonmcp/simon"],
      "env": {
        "SIMON_PROJECT_PATH": "/path/to/your/project"
      }
    }
  }
}

Usage

Available Tools

claude Tool

Execute development commands using natural language with optimized parameter structure:

// Basic usage
claude({
  task_description: "写一个Python函数计算圆的面积"
})

// With context and specifications
claude({
  task_description: "开发一个贪吃蛇游戏",
  context: "这是给中学生用的教学代码，需要简单易懂",
  spec: {
    language: "JavaScript",
    format_requirements: "使用ES6语法和Canvas API",
    project_path: "/path/to/project"
  }
})

// With additional flags
claude({
  task_description: "分析代码质量",
  spec: {
    project_path: "/path/to/project",
    additional_flags: ["--verbose"]
  }
})

// Complex example
claude({
  task_description: "修复用户登录功能的bug",
  context: "用户反馈登录时出现500错误",
  spec: {
    language: "TypeScript",
    format_requirements: "添加类型注释和错误处理",
    project_path: "/path/to/auth/module"
  }
})

Parameter Description

| Field | Type | Required | Description | |-------|------|----------|-------------| | task_description | string | ✅ | Clear description of what code you want, e.g., "写一个Python函数计算圆的面积" | | context | string | ❌ | Additional background information, e.g., "这是给中学生用的教学代码" | | spec | object | ❌ | Specific requirements: language/format/standards | | spec.language | string | ❌ | Programming language, e.g., "Python", "JavaScript", "TypeScript" | | spec.format_requirements | string | ❌ | Format requirements, e.g., "添加类型注释", "使用ES6语法" | | spec.project_path | string | ❌ | Project path (optional) | | spec.additional_flags | array | ❌ | Additional flag parameters |

help Tool

Get help information:

// General help
help()

// Specific topics
help({ topic: "commands" })
help({ topic: "flags" })
help({ topic: "examples" })
help({ topic: "config" })

Supported Command Types

1. Refactor (重构) - Code refactoring and improvement
2. Develop (开发) - Create new features or applications
3. Analyze (分析) - Project analysis and insights
4. Fix (修复) - Bug fixing and debugging
5. Test (测试) - Testing and quality assurance
6. Document (文档) - Documentation generation
7. Optimize (优化) - Performance optimization
8. Deploy (部署) - Deployment and distribution

Command Flags

• -p - Project mode (work within project context)
• --project <path> - Specify project path
• --context <json> - Additional context information
• --dry-run - Show what would be done without executing
• --verbose - Enable verbose output

Configuration

Environment Variables

• SIMON_PROJECT_PATH - Default project path for operations
• SIMON_ALLOWED_PATHS - Comma-separated list of allowed paths

Configuration File

Create a simon-config.json file:

{
  "allowedPaths": ["/path/to/project1", "/path/to/project2"],
  "allowedOperations": ["read", "write", "create"],
  "allowedFileExtensions": [".js", ".ts", ".py", ".md"],
  "maxFileSize": 10485760,
  "defaultProjectPath": "/default/project/path"
}

Security Features

• Path Sandboxing: Operations restricted to configured allowed paths
• File Type Filtering: Only specified file extensions are accessible
• Size Limits: Files larger than configured limit are rejected
• Operation Permissions: Read/write/create operations can be controlled

Examples

Basic Commands

# Refactor current project
claude "对目前项目代码进行重构" -p

# Develop a new game
claude "开发一个贪吃蛇游戏" --project "/path/to/game/project"

# Analyze code quality
claude "分析项目的代码质量和结构" --verbose

# Fix a specific bug
claude "修复用户登录功能的bug" --context '{"module": "auth"}'

Integration with MCP Clients

Once configured, you can use natural language commands in your MCP-enabled AI assistant:

• "Please refactor the current project code"
• "Develop a snake game for me"
• "Analyze the project structure and suggest improvements"
• "Help me fix the login bug"

Development

Building from Source

git clone <repository>
cd simon-mcp
npm install
npm run build

Running in Development

npm run dev

Testing

# Test with MCP inspector
npx @modelcontextprotocol/inspector node dist/index.js

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.