AI Agent Security 2026: Harden MCP & Block Prompt Injection

About this course

“This course contains the use of artificial intelligence.”

AI agent security is a boundary problem. Once a language model can call tools, read files, send requests and remember what it saw, prompt injection and over-privileged tools stop being theory. This hands-on course shows you how to secure AI agents and MCP servers layer by layer, on a local lab bench you build yourself. No GPU and no paid AI API are needed: the labs run on CPU with a small open model.

What you will build and testA local agent bench: Ollama, a small tool-calling model and a Python agent loop that logs every step

Your own MCP server, hardened and validated in both directionsOPA/Rego policies that decide every tool call, including the call a prompt injection tries to make

Sandboxed tool runtimes, short-lived audience-bound tokens and human approval gates

Audit logs, detection rules, a kill-switch drill and a control map to the OWASP Top 10 for Agentic Applications (ASI01–ASI10) and the NIST AI RMF

You start by building that bench: Ollama in a container bound to localhost, a small tool-calling model chosen for size, speed and licence, and a minimal agent loop in Python that logs every step. From the first tool call you see how the model chooses arguments, including invented ones, and why validation has to start at the tool boundary.

Next you read the Model Context Protocol from the defender's seat. You write your own MCP server with the Python SDK, list exactly what the model will see, learn why tool annotations are hints rather than controls, and validate both tool inputs and structured results. Least privilege follows: an open-ended shell tool is replaced by narrow functions, file paths are confined correctly, and read and write access are split across separate servers.

Every tool call then passes through one decision point.

Read the full description on Udemy on Udemy (opens in a new tab)

Prices, discounts and availability are set by Udemy. We may earn a commission when you purchase through links on this site.