MCP Applications & RAG over MCP

Practise and Check

Deliverable

Shared retrieval: connect the RAG-over-MCP server to two different clients (a LangGraph agent and Claude Desktop or a local agent) and compare answers.

Assessment

Answer the five Interview questions without notes, then run the lab.

Interview probe

Larkspur's LangGraph support assistant loads the handbook server with `MultiServerMCPClient.get_tools()`. Latency per answer grows with the number of retrieval calls, and the server's startup log line appears once per call. What is happening, and what do you change?

Common mistakes

• Printing banners to stdout on a stdio server instead of stderr
• Letting the model do the arithmetic while the servers only move cells
• Trusting similarity_score from a squared-L2 collection as cosine

Related Modules

• T12.1MCP Foundations
• T12.2Building MCP Servers & Clients
• T12.4A2A: Agent-to-Agent Protocol