// Learning paths

Follow a path, not a pile of labs.

Structured, hands-on journeys that take you from your first prompt injection to compromising whole agent systems — each step builds on the last.

Prefer one topic at a time? Browse modules →

Learning paths

Learning path

AI Red Teaming

From your first prompt injection to compromising a multi-agent system.

Hard0 labs · 4 modules
Learning path

Prompt Injection Fundamentals

Master the number-one LLM vulnerability, start to finish.

Easy0 labs · 2 modules
Learning path

Model & Data Attacks

Attack the model itself — poison its training, steal its behaviour, breach its privacy.

Medium0 labs · 3 modules
Learning path

Agentic Exploitation

Turn a tool-calling agent's own capabilities into remote code execution.

Hard0 labs · 3 modules
Learning path

AI Supply Chain & MLOps

Poison the pipeline once and ship your payload to every model downstream.

Hard0 labs · 3 modules