Before you can attack an AI agent you have to understand what turns a chatbot into one. This module covers the components of a single agent, the reasoning loop, and a map of every place untrusted text becomes an instruction.
The noisy half of recon: discovering AI services on the network, fingerprinting the backend model by behavior, and mapping a RAG pipeline with direct probes. The final part of the reconnaissance module.
How much of an AI stack you can map before sending a single adversarial request: fingerprinting from HTTP responses, and mining public code and artifacts for models, endpoints, and secrets.
Before you can attack an AI system you have to understand what it's made of. This module covers the anatomy of an AI-enabled stack and a taxonomy to organize everything you enumerate.
An independent, hands-on study series for the OSAI (OffSec AI Red Teamer) exam. Technical notes, real commands, and the offensive-AI concepts you need, one module at a time.
Skip reading about how Gandalf-style guardrails work and build one yourself. A free, offline Python lab using Ollama that implements all eight defense patterns from this series, ready for you to attack.
The final Gandalf levels stack every defense and then harden them with data from a million real attacks. They're beatable too, and that's the most important lesson in the whole series.