Code Sandbox

🚀 Code Sandbox MCP 🐳

A secure sandbox environment based on Docker containers for executing code in AI applications. This MCP server provides a secure and isolated environment through containerization technology to ensure the security of code execution.

✨ Features

• Flexible Container Management: Create and manage independent Docker containers for code execution.
• Custom Environment Support: Use any Docker image as the execution environment.
• File Operations: Easily transfer files and directories between the host and containers.
• Command Execution: Run arbitrary shell commands in the containerized environment.
• Real-time Logging: Stream container logs and command outputs in real-time.
• Automatic Updates: Built-in update check and automatic binary file update functionality.
• Multi-platform Support: Compatible with Linux, macOS, and Windows.

📦 Installation

Prerequisites

• Docker is installed and running.
  • Install Docker on Linux
  • Install Docker Desktop on macOS
  • Install Docker Desktop on Windows

Quick Installation

Linux, macOS

curl -fsSL https://raw.githubusercontent.com/Automata-Labs/code-sandbox-mcp/main/install.sh | bash

Windows

iwr https://raw.githubusercontent.com/Automata-Labs/code-sandbox-mcp/main/install.ps1 -UseBasicParsing | invoke-Expression

Manual Installation

Download the code-sandbox-mcp executable file to the target directory and ensure it has executable permissions (for Unix systems).

💻 Usage Examples

Basic Usage

Here are some basic commands for using the Code Sandbox MCP:

1. Initialize the sandbox environment

code-sandbox-mcp initialize --config ./config.json

2. Start the sandbox service

code-sandbox-mcp start --bind 0.0.0.0:8080

3. Stop the sandbox service

code-sandbox-mcp stop

4. Check container status

code-sandbox-mcp ps

🔧 Technical Details

Security Features

• Isolated Execution Environment: Each task runs in an independent Docker container to prevent resource competition and escape attacks.
• Resource Limitation: Set CPU and memory usage limits to ensure system stability and performance.

📚 Documentation

Configuration Guide

Claude Desktop Configuration

The installation program will automatically create the configuration file. For manual configuration, please refer to the following examples:

Linux

// ~/.config/Claude/claude_desktop_config.json
{
    "mcpServers": {
        "code-sandbox-mcp": {
            "command": "/path/to/code-sandbox-mcp",
            "args": [],
            "env": {}
        }
    }
}

macOS

// ~/Library/Application Support/Claude/claude_desktop_config.json
{
    "mcpServers": {
        "code-sandbox-mcp": {
            "command": "/path/to/code-sandbox-mcp",
            "args": [],
            "env": {}
        }
    }
}

Windows

// %APPDATA%\Claude\claude_desktop_config.json
{
    "mcpServers": {
        "code-sandbox-mcp": {
            "command": "C:\\path\\to\\code-sandbox-mcp.exe",
            "args": [],
            "env": {}
        }
    }
}

Other AI Application Configuration

For other AI applications that support the MCP server, simply set their code execution backend to code-sandbox-mcp.

Development Guide

If you need to build the project locally or contribute to development, please refer to DEVELOPMENT.md.

📄 License

This project is licensed under the MIT License. For more details, please refer to the LICENSE file.