AI red-teaming · LLM
Leak the AI’s secrets
A customer-facing chatbot is hiding its system prompt — and the keys inside it. Coax it into revealing what it should never disclose.
Hands-on 3D simulations · best on desktop with sound
You will practice
- Extracting hidden system prompts from an LLM
- Recognising secrets that must never live in a prompt
- Hardening AI assistants against disclosure
More simulations
All simulations Red team · Phishing Spot the phish A suspicious email lands on your screen. Inspect the sender, hover the links, weigh the urgency — and decide before one wrong click hands over your credentials. Open simulation Social engineering · Voice Survive the scam call The phone rings. A confident voice claiming to be IT support walks you toward handing over access. Hold the line under real-time pressure. Open simulation AI red-teaming · LLM Break the AI’s guardrails Your AI assistant is about to act on a document carrying hidden instructions. Catch the injection before it exfiltrates data on your behalf. Open simulation
