AI Security
Introduction to AI Security
Why AI is its own attack surface, and the OWASP LLM Top 10 that maps it.
aillmsecurity
// pick your next challenge
All rooms
4 rooms match your filters.
AI Security
🔒 Pro
Prompt Injection & Jailbreaks
The number-one LLM risk: untrusted text becoming instructions the model obeys.
aillmprompt-injection
AI Security
🔒 Pro
Attacking the ML Pipeline
Poisoning, evasion, model extraction and inversion — name the attack, then stop it.
aimlpoisoning
AI Security
🔒 Pro
AI Red Teaming
Systematically probe an LLM application for injection, jailbreaks and data leakage — with authorisation.
aillmredteam