You must expose an internal procedures API and a semantic search index to multiple agents without hard-coding HTTP clients in every agent. The platform standard is Model Context Protocol (MCP). What design best integrates knowledge for multi-agent consumption?
Select an answer to reveal the explanation.
Short Explanation
You don't want every agent to carry its own HTTP client or secrets. Build MCP servers, such as Azure Functions or APIM-fronted backends, that expose search and procedure tools, then connect agents as MCP clients with scoped tool catalogs. That's the clean way to integrate knowledge without hard-coding clients.
Full Explanation
Correct design: build MCP servers, such as Azure Functions or APIM-fronted backends, that expose search and procedure tools, then connect agents as MCP clients with scoped tool catalogs. Knowledge integration includes search, RAG, MCP-available sources, and semantic search. MCP servers expose tools and knowledge; agents act as clients with scoped catalogs. Azure Functions, Logic Apps, and APIM are common hosts or front doors.
Embedding API keys in prompts and inventing REST URLs violates security and reliability. Giving only the orchestrator raw database credentials and banning specialist tools breaks least privilege and specialist design. Nightly CSV email dumps pasted into chat are not scalable multi-agent knowledge integration.
Exam tip: Prefer MCP or an equivalent tool protocol over per-agent hard-coded clients.