Frequently Asked Questions (FAQ)

Q: What is an MCP Server?

A: MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs. An MCP server acts as a bridge, allowing AI models to access and interact with external data sources and tools in a controlled and standardized manner.

Q: Why use a Node.js Sandbox MCP Server?

A: It provides a secure and isolated environment for executing JavaScript code, especially useful for AI agent development where code might interact with sensitive data or external APIs. It allows for dynamic dependency installation and ephemeral execution, ensuring clean and reproducible runs.

Q: What are the prerequisites for using this MCP Server?

A: You need to have Docker installed and running on your machine, as the server relies on Docker containers for creating isolated environments.

Q: How do I install npm dependencies within the sandbox?

A: When using the run_js or run_js_ephemeral tools, you can specify an array of dependencies with their names and versions. The server will automatically install these dependencies before executing your code.

Q: What is the difference between run_js_ephemeral and run_js?

A: run_js_ephemeral runs a one-off JavaScript script in a brand-new, disposable container. run_js requires a pre-initialized container (using sandbox_initialize) and allows you to run multiple scripts within the same container, potentially reusing dependencies and maintaining state.

Q: What is detached mode and when should I use it?

A: Detached mode, enabled by setting listenOnPort in run_js, keeps the container alive after script execution. It’s useful for long-running processes like servers or services that need to remain active and accessible.

Q: How can I retrieve files generated by my JavaScript code?

A: Use the run_js_ephemeral tool. If your code saves files in the current directory during execution, these files will be automatically returned. Images are returned as image content, and other files as resource content.

Q: How does the UBOS platform integrate with the MCP Server?

A: The UBOS platform helps orchestrate AI Agents, connect them with your enterprise data, build custom AI Agents with your LLM model, and create Multi-Agent Systems, making the MCP Server a crucial component for secure and controlled code execution within these agents.

Q: Can I limit the resources used by the containers?

A: Yes, containers run with controlled CPU and memory limits to prevent resource exhaustion and ensure fair allocation.

Q: Where can I find more examples and usage tips?

A: Check out the USE_CASES.md file in the GitHub repository for a curated list of prompts, examples, and creative experiments.