Sandbox FAQs

Question 1

What is Sandbox and what does it do?

Accepted Answer

Sandbox is a Model Context Protocol (MCP) server that allows Large Language Models (LLMs) to run code in secure, isolated Docker containers. It prevents potentially harmful code from directly impacting your system.

Question 2

How does Sandbox integrate with Large Language Models (LLMs)?

Accepted Answer

Sandbox integrates with LLMs using the Model Context Protocol (MCP). This allows LLMs to send code to Sandbox for execution and receive the results, all within a secure and controlled environment.

Question 3

What sandbox environments are supported?

Accepted Answer

Sandbox supports multiple environments including shell, Python, JavaScript, Go, and network-tools. You can also create custom sandboxes using Dockerfiles and configuration files.

Question 4

Why use Sandbox instead of running code directly?

Accepted Answer

Directly running LLM-generated code can pose security risks. Sandbox provides strict security and network constraints within its Docker containers, protecting your system from unintended consequences.

Question 5

How do I create my own custom sandbox environment?

Accepted Answer

You can create custom sandboxes by adding a new directory in the `sandboxes` directory with your sandbox name, and adding a `Dockerfile` and `config.json` to it. See the existing sandboxes for examples.

Sandbox

Sandbox

Key Features

Use Cases

Key Features

Use Cases