Tools and MCP (LangGraph sample)

What you see when MCP is enabled

DataRobot can attach extra tools (including from an MCP deployment) when the agent runs. In the graph sample, those appear alongside tools you define in Python. If you list only your own tools and ignore the injected list, MCP tools will not be available to the model.

Practical rule: merge platform tools with yours wherever you bind tools to the graph (the e2e myagent.py shows the intended pattern).

Configuration outside the repo

MCP server URL and credentials come from your deployment / environment (MCP deployment id, external MCP URL, etc.), not from the LangGraph workflow.yaml in the minimal sample. The NAT workflow example adds function_groups in YAML instead; see nat/mcp.md.

Automated tests

e2e-tests/dragent_tests/test_mcp.py exercises MCP tool calls when MCP_DEPLOYMENT_ID and a tool-capable agent are configured.