About this session
Your AI assistant is only as useful as the tools it can reach. The Model Context Protocol (MCP) is the standard that finally makes those tools portable: write a capability once, and it plugs into any MCP-aware host — your own agent framework, Claude Code, Codex, OpenCode — with no per-app glue code.
In this hands-on workshop we'll start from the agentic loop (framework → LLM → tool call → local execution → result) and see exactly where MCP fits. We'll cover what MCP actually is, how it compares to just handing an agent a CLI, and what separates a tool an LLM uses well from one it fumbles.
Then we build a real MCP server from scratch in Python, test it with no LLM in the loop, and wire the same unchanged server into four different AI assistants live. We'll close on the part everyone skips: the security implications of running tools that act on your behalf.
What you’ll take away
- A clear mental model of where MCP fits in the agentic loop
- A working MCP server built from scratch in Python
- Patterns that separate LLM-friendly tools from ones models fumble
- The security implications of running tools that act on your behalf
