Code Sandbox

What is Code Sandbox about?

Creates and manages isolated Docker containers that serve as execution environments for arbitrary code. It abstracts container lifecycle, file transfer, command execution, and log streaming, giving AI services a reliable way to run user‑provided scripts without compromising host security.

How to use Code Sandbox?

1. Install – run the installer script for your platform (Linux/macOS: curl … | bash; Windows: PowerShell irm … | iex). The installer checks Docker, downloads the binary, and creates a config file.
2. Initialize a sandbox – call the sandbox_initialize tool, optionally specifying a Docker image (default python:3.12-slim-bookworm). It returns a container_id.
3. Transfer files – use copy_project, copy_file, or write_file to place code and resources inside the container.
4. Execute commands – invoke sandbox_exec with an array of shell commands (e.g., install dependencies, run scripts).
5. Monitor output – fetch real‑time logs via the containers://{id}/logs resource or read stdout/stderr streams.
6. Cleanup – stop and remove the container with sandbox_stop.

Key Features

• Flexible container management with any Docker image
• Seamless file and directory transfer between host and container
• Arbitrary command execution inside the sandbox
• Real‑time logging and output streaming
• Automatic update checking and binary self‑update
• Cross‑platform support (Linux, macOS, Windows)
• Resource limits enforced by Docker for security

Use Cases

• Running untrusted user code in AI coding assistants
• Executing data‑processing scripts in automated pipelines
• Providing isolated runtimes for code‑generation models
• Testing code snippets safely within a CI/CD workflow
• Offering sandboxed environments for educational platforms

FAQ

Q: Do I need Docker installed? A: Yes, Docker must be running on the host; the installer validates this.

Q: Can I use a custom Docker image? A: Absolutely – pass the image parameter to sandbox_initialize with any valid image name.

Q: How are resources limited? A: Limits are applied via Docker container constraints (CPU, memory) that you can configure when building your image.

Q: Is there automatic cleanup? A: Containers persist until you call sandbox_stop; the stop command gracefully shuts down and removes the container and its volumes.

Q: How do I access logs programmatically? A: Use the dynamic resource containers://{id}/logs, which returns the full log as plain text.

Code Sandbox MCP 🐳

A secure sandbox environment for executing code within Docker containers. This MCP server provides AI applications with a safe and isolated environment for running code while maintaining security through containerization.

🌟 Features

• Flexible Container Management: Create and manage isolated Docker containers for code execution
• Custom Environment Support: Use any Docker image as your execution environment
• File Operations: Easy file and directory transfer between host and containers
• Command Execution: Run any shell commands within the containerized environment
• Real-time Logging: Stream container logs and command output in real-time
• Auto-Updates: Built-in update checking and automatic binary updates
• Multi-Platform: Supports Linux, macOS, and Windows

🚀 Installation

Prerequisites

• Docker installed and running
  • Install Docker for Linux
  • Install Docker Desktop for macOS
  • Install Docker Desktop for Windows

Quick Install

Linux, MacOS

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

Windows

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

The installer will:

1. Check for Docker installation
2. Download the appropriate binary for your system
3. Create necessary configuration files

Manual Installation

1. Download the latest release for your platform from the releases page
2. Place the binary in a directory in your PATH
3. Make it executable (Unix-like systems only):

   chmod +x code-sandbox-mcp
   

🛠️ Available Tools

sandbox_initialize

Initialize a new compute environment for code execution. Creates a container based on the specified Docker image.

Parameters:

• image (string, optional): Docker image to use as the base environment
  • Default: 'python:3.12-slim-bookworm'

Returns:

• container_id that can be used with other tools to interact with this environment

copy_project

Copy a directory to the sandboxed filesystem.

Parameters:

• container_id (string, required): ID of the container returned from the initialize call
• local_src_dir (string, required): Path to a directory in the local file system
• dest_dir (string, optional): Path to save the src directory in the sandbox environment

write_file

Write a file to the sandboxed filesystem.

Parameters:

• container_id (string, required): ID of the container returned from the initialize call
• file_name (string, required): Name of the file to create
• file_contents (string, required): Contents to write to the file
• dest_dir (string, optional): Directory to create the file in (Default: ${WORKDIR})

sandbox_exec

Execute commands in the sandboxed environment.

Parameters:

• container_id (string, required): ID of the container returned from the initialize call
• commands (array, required): List of command(s) to run in the sandboxed environment
  • Example: ["apt-get update", "pip install numpy", "python script.py"]

copy_file

Copy a single file to the sandboxed filesystem.

Parameters:

• container_id (string, required): ID of the container returned from the initialize call
• local_src_file (string, required): Path to a file in the local file system
• dest_path (string, optional): Path to save the file in the sandbox environment

sandbox_stop

Stop and remove a running container sandbox.

Parameters:

• container_id (string, required): ID of the container to stop and remove

Description: Gracefully stops the specified container with a 10-second timeout and removes it along with its volumes.

Container Logs Resource

A dynamic resource that provides access to container logs.

Resource Path: containers://{id}/logs
MIME Type: text/plain
Description: Returns all container logs from the specified container as a single text resource.

🔐 Security Features

• Isolated execution environment using Docker containers
• Resource limitations through Docker container constraints
• Separate stdout and stderr streams

🔧 Configuration

Claude Desktop

The installer automatically creates the configuration file. If you need to manually configure it:

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 Applications

For other AI applications that support MCP servers, configure them to use the code-sandbox-mcp binary as their code execution backend.

🛠️ Development

If you want to build the project locally or contribute to its development, see DEVELOPMENT.md.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.