sandbox_initializeLABSCO SUMMARY
The image is yours to pick, so the sandbox matches the code it is running.
Initialisation takes any Docker image and hands back a container id the later calls address, so Python code runs in a Python image and Go code in a Go one without a second server or a rebuilt environment.
What it is
A code execution sandbox backed by Docker. Each session creates a container from an image you name, runs shell commands inside it, and stops it when you are done.
What you get
- sandbox_initialize — creates a container from a Docker image you choose, and returns the container id every other call takes
- sandbox_exec — runs shell commands inside that container, with output streamed back as it happens
- copy_project — copies a whole directory into the sandbox filesystem
- copy_file and write_file — a single file copied in from your machine, or written directly
- sandbox_stop — ends the container
Requirements
Docker installed and running. The server is a single binary — the vendor's install script places it on Linux, macOS and Windows, or take it from the releases page.
Setup effort
One command — curl -fsSL https://raw.githubusercontent.com/Automata-Labs-team/code-sandbox-mcp/main/install.sh | bash
USE CASES
01
Letting an assistant actually run the code it wrote
Start a container, write the file into it and execute — the model sees real output instead of predicting it.
02
Testing a change against a real project
Copy the whole working directory in, then run the test command inside the container rather than on your machine.
03
Trying a dependency without installing it locally
Install and use it inside a throwaway container, and stop the container when the question is answered.
04
Reproducing a bug on a different runtime version
Pick the image that matches the version in question, copy the project in and run it there.
6 tools
Listed in project docs. Your client may show a different list.
copy_projectwrite_filesandbox_execcopy_filesandbox_stopBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
Install and start Docker (prerequisite)
Run the install script: curl -fsSL https://raw.githubusercontent.com/Automata-Labs-team/code-sandbox-mcp/main/install.sh | bash (Windows PowerShell: irm https://raw.githubusercontent.com/Automata-Labs-team/code-sandbox-mcp/main/install.ps1 | iex)
The installer checks Docker, downloads the binary for your platform and creates the config file — no manual config edit is normally needed
MCP Clients
Any MCP client
ANY MCP CLIENT · JSON
{
"mcpServers": {
"code-sandbox-mcp": {
"command": "<YOUR_PATH_TO_CODE_SANDBOX_MCP>",
"args": [],
"env": {}
}
}
}
