Labs
Put the ideas into practice. Explore AI attack surfaces and the controls that protect them.
AI red team training
Project 2501
A hands-on training platform for investigating and defending AI systems.
Work through guided labs and black-box exercises covering prompt injection, AI infrastructure, retrieval systems and agent/MCP security. Practise investigating a weakness, recording evidence and evaluating the defensive controls that should stop it.
Each exercise connects the attack path to practical security testing and evidence you can explain.