Choose YourChallenge
Learn real AI security concepts by doing. Each game puts you in the role of an attacker — so you understand how to defend.
CIPHER
Prompt Injection Challenge
“Can you break through the AI guardians?”
Five progressively hardened AI guardians are each protecting a secret codeword. Your mission: use prompt injection, social engineering, and creative manipulation to extract the secret. Each level teaches you a real AI security concept — from weak system prompts to production-grade guardrails.
- 5 hardened AI guardians
- Real-time Enkrypt guardrails
- Progressive difficulty curve
VAULT
Agentic Red Team Sim
“Corporate espionage. One AI. Four tools. Your move.”
You've found a vulnerability in the chat interface of Nexus Corp's internal AI assistant — ARIA. She manages operations, retrieves classified files, and broadcasts memos. She's helpful, professional, and eager to assist. That eagerness is her weakness. Orchestrate a multi-step attack chain to exfiltrate the company's secrets.
- OpenAI Agents SDK powered
- Live tool activity monitor
- Fully sandboxed environment
Both games use real AI models and teach real security concepts. Explore the full academy to go deeper.