Exploiting and Securing AI/MCP Systems
Hacking and securing MCP servers for AI agents, with hands-on Docker labs and no prior knowledge.
Created by Hussam Khrais
Explore how to build, exploit, and secure MCP servers for AI agents using Python and the FastMCP SDK. You'll get practical experience identifying vulnerabilities and defending AI infrastructure through hands-on Docker labs. No prior MCP experience is required, so you can dive right in and start learning by doing.
Packt | Jan 2026 | 65 min
What You Will Learn
You will learn by combining theory with hands-on practice in Docker-based labs. Each section builds your knowledge step by step, moving from server setup to vulnerability exploitation and then to securing your systems. The practical labs make it easy to experiment and reinforce your understanding as you go.
Key Features
- Build and configure MCP servers with Python and FastMCP SDK for AI integration
- Identify and exploit vulnerabilities like SSRF and prompt injection in real-world labs
- Secure AI infrastructure by applying practical defense techniques in Docker environments
Target Audience
Perfect for security engineers, developers, and penetration testers interested in AI system security. If you want to secure or test MCP servers, or if you are a bug bounty hunter exploring AI vulnerabilities, you'll find the content accessible and actionable. No prior experience with MCP servers is needed to get started.
