Code Sandbox MCP Server

Overview

Code Sandbox creates secure, isolated environments for executing code using Docker containers. This tool is particularly valuable for AI applications that need to run untrusted or experimental code safely without compromising the host system.

Installation

Prerequisites

Before installing Code Sandbox, ensure you have Docker installed and running on your system:

• Docker for Linux
• Docker Desktop for macOS
• Docker Desktop for Windows

Installation Methods

Automatic Installation

For Linux and macOS:

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

For Windows (PowerShell):

irm https://raw.githubusercontent.com/Automata-Labs-team/code-sandbox-mcp/main/install.ps1 | iex

Manual Installation

1. Download the latest release for your platform from the GitHub releases page
2. Place the binary in a directory included in your PATH
3. For Unix-like systems, make the binary executable:

   chmod +x code-sandbox-mcp
   

Usage

Code Sandbox provides a suite of tools for managing containerized environments:

1. Creating a Sandbox: Initialize a new sandbox environment using sandbox_initialize with your preferred Docker image
2. File Management: Transfer files and directories between your host system and containers
3. Command Execution: Run commands within the isolated environment using execute_command
4. Environment Management: Manage container lifecycle with tools like sandbox_stop

Example Workflow

1. Initialize a Python sandbox:

   # Tool: sandbox_initialize
   # Parameters: {"image": "python:3.12-slim-bookworm"}
   

2. Copy a project into the sandbox:

   # Tool: copy_project
   # Parameters: {"container_id": "your_container_id", "source_path": "./my_project", "destination_path": "/app"}
   

3. Execute code in the sandbox:

   # Tool: execute_command
   # Parameters: {"container_id": "your_container_id", "command": "cd /app && python main.py"}
   

4. Retrieve output files:

   # Tool: copy_file_from_sandbox
   # Parameters: {"container_id": "your_container_id", "source_path": "/app/output.txt", "destination_path": "./results/output.txt"}
   

5. Stop the sandbox when finished:

   # Tool: sandbox_stop
   # Parameters: {"container_id": "your_container_id"}
   

Security Considerations

While Code Sandbox provides isolation through Docker containers, be aware of potential security implications:

• The sandbox requires Docker access, which has root-equivalent permissions
• Resource limits should be configured to prevent container-based DoS attacks
• File system access between host and container should be carefully managed