Code Sandbox MCP server for AI agents

Code Sandbox MCP is a secure sandbox environment for executing code within Docker containers. It provides AI applications with a safe and isolated environment for running code while maintaining security through containerization, offering features like flexible container management, custom environment support, and real-time logging.

Installation

Prerequisites

Before installing Code Sandbox MCP, ensure you have Docker installed and running:

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:

Check for Docker installation
Download the appropriate binary for your system
Create necessary configuration files

Manual Installation

Download the latest release for your platform from the releases page
Place the binary in a directory in your PATH
Make it executable (Unix-like systems only):
```
chmod +x code-sandbox-mcp
```

Available Tools

sandbox_initialize

Initialize a new compute environment for code execution.

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.

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.

How to add this MCP server to Cursor

There are two ways to add an MCP server to Cursor. The most common way is to add the server globally in the ~/.cursor/mcp.json file so that it is available in all of your projects.

If you only need the server in a single project, you can add it to the project instead by creating or adding it to the .cursor/mcp.json file.

Adding an MCP server to Cursor globally

To add a global MCP server go to Cursor Settings > MCP and click "Add new global MCP server".

When you click that button the ~/.cursor/mcp.json file will be opened and you can add your server like this:

{
    "mcpServers": {
        "cursor-rules-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "cursor-rules-mcp"
            ]
        }
    }
}

Adding an MCP server to a project

To add an MCP server to a project you can create a new .cursor/mcp.json file or add it to the existing one. This will look exactly the same as the global MCP server example above.

How to use the MCP server

Once the server is installed, you might need to head back to Settings > MCP and click the refresh button.

The Cursor agent will then be able to see the available tools the added MCP server has available and will call them when it needs to.

You can also explictly ask the agent to use the tool by mentioning the tool name and describing what the function does.