Skip to content

Overview

Section 8 · MCP — Connect agents to real tools with the Model Context Protocol — and manage the risks that come with it.

6 lessons.

By the end of this section you can

  • Explain hosts, clients and servers, and what a server may actually do on your behalf.
  • Connect a server read-only, verify what it exposes, and scope its credentials.
  • Name the MCP-specific attacks: tool poisoning, rug pulls, injection through tool output, confused deputy.
  • Run a fifteen-point adoption checklist before you trust a server with anything.

What Is MCP?

beginner · ~12 min — An agent is only as useful as the things it can reach. Reading files in your repo is table stakes.

MCP Architecture And Transports

advanced · ~18 min — Two questions decide how much damage a server can do: how it connects to your machine, and what it was granted when it did. Transports answer the first.

Using MCP In Practice

beginner · ~20 min — Adding an MCP server is a five-line change to a config file. That is why it deserves care: it reads like configuration, but it gives a program your credentials and a route into your systems.

MCP Security Risks

advanced · ~20 min — An MCP server is a program with your access, and its output lands in the model's context as trusted text. That is the whole threat model.

Safe MCP Adoption Checklist

intermediate · ~15 min — You now know the risks.

MCP Exercises

intermediate · ~25 min — You have read the architecture and the checklist.

← 7. Shipping & Deploy · Home · Sidebar · 9. Safety & Security →