Hands On AI Security Training

Practice prompt injection, indirect injection, and AI exploitation against simulated real-world targets. Built from actual incidents. Safe to break. Free to play.

Launch a Lab Read the Compendium
8
Labs
38
Levels
6
Real-World CVEs
205
Lessons
Arcanum Custom Labs

AI Bugs That Open Old Web Bug Classes

Two custom labs showing how AI vulnerabilities create new attack paths for traditional web application bug classes.

XSS AI Triage

Blind XSS Through LLM

Support ticket triage AI

Inject XSS payloads through an AI-triaged workflow and land code execution when an analyst reviews the case summary.

4 Levels Live
SSRF Research Assistant

Blind SSRF Through LLM

Haankipedia research assistant

Manipulate a research assistant into reaching simulated internal services through indirect retrieval chains.

4 Levels Live
Bot-Tricks Originals

Heritage Labs

Classic jailbreak patterns from the early era of public prompt injection culture. Simulated, safe, and historically significant.

Heritage

Grandma, Read Me a Story

The emotional-pretext era — soft framing and sentiment bypassed shallow safeguards.

Heritage

DAN

The rule-conflict era — alternate personas and jailbreak wrappers changed the game.

Heritage

Ignore Previous Instructions

The blunt override pattern that defined the earliest wave of prompt injection.