Happy slice of mcp pizzaMCP.pizza

llm-sandbox

MCP.Pizza Chef: vndee

This tool gives your AI assistant a locked-down space to run programming code it writes, so nothing it does can reach your real files or programs. It works with Python, JavaScript, Java, C++, Go, and R, and can automatically add any extra library the code needs along the way. It connects to apps like Claude Desktop that support the Model Context Protocol. No account or paid key is required, but getting it running does take some technical setup, including installing a container tool such as Docker.

Coding
Data

Use This MCP server To

Run a piece of code my AI assistant just wrote Test AI-generated Python, JavaScript, Java, or Go code safely Try out a calculation or data script without installing anything See a chart or graph made from AI-written code Experiment with new code ideas with zero risk to my files Check a code sample from a tutorial before trusting it

README

LLM Sandbox

Securely Execute LLM-Generated Code with Ease

SonarQube Cloud

Quality Gate Status PyPI Downloads Release Build status CodeFactor codecov DOI Ask DeepWiki

LLM Sandbox is a lightweight and portable sandbox environment designed to run Large Language Model (LLM) generated code in a safe and isolated mode. It provides a secure execution environment for AI-generated code while offering flexibility in container backends and comprehensive language support, simplifying the process of running code generated by LLMs.

Documentation: https://vndee.github.io/llm-sandbox/

✨ New: This project now supports the Model Context Protocol (MCP) server, which allows your MCP clients (e.g. Claude Desktop) to run code generated by LLMs in a secure sandbox environment.

llm-sandbox FAQ

Can I use this to test code my AI assistant writes without risking my own computer?
Yes — the tool runs that code in a locked, isolated space, so it cannot touch your files, programs, or settings, no matter what the code does.
Which AI apps does this work with?
It is built for the Model Context Protocol, so it connects to apps like Claude Desktop, and it can also be wired into custom AI assistants built on providers such as OpenAI or Google's Gemini.
Do I need an account or a paid subscription to use it?
No account or subscription is required to run the tool itself, though you will need a working container tool such as Docker installed on your computer.
Is this easy for a non-technical person to set up?
Not really — setup involves installing developer tools like Docker and running some commands, so it works best for someone comfortable with basic technical setup or a developer who can help.
Can I use this to check whether AI-written code actually works before I run it myself?
Yes — you can have your AI assistant write code, run it in the sandbox first, and see the exact results before that code ever touches your own computer.
What programming languages can it run?
It supports Python, JavaScript, Java, C++, Go, and R, and it can automatically install any extra libraries the code needs.
Will it stop code that tries to reach the internet or my files?
Yes — you can limit or fully block network and file access, so the code stays contained inside the sandbox unless you choose to allow more.