feat: MCP Code Sandbox #166
Description
Many of the agentic works need a code to be executed, a few example is
- WO Agent
- Rule Logic Agent
- NL2ML application
- TSFM fine-tuning (also a Jupyter notebook)
- Data Quality Agent
- MetaAgent
Given that we are rebasing AssetOpsBench in the MCP-compliant environment, we would like to provide a Sandbox as part of the AssetOpsBench MCP echo system.
I have finded few entry point
https://navendu.me/posts/sandbox-mcp/
https://github.com/datalayer/jupyter-mcp-server
https://github.com/langchain-ai/langchain-sandbox
What shall we provide?