7 hands-on practice environments for AI red teaming and prompt injection, with difficulty ratings from beginner to advanced
Hands-on prompt injection labs with real LLMs, a Gauntlet CTF, and a layered Context Trace where protected values may be redacted.
Wargame-style challenge where players must trick an LLM into revealing secret passwords across progressively harder levels.
Intentionally vulnerable AI agent designed for practicing LLM exploitation techniques.
Collection of 9 labs covering path traversal, RCE, prompt injection, and supply-chain attacks on MCP.
Vulnerable recruitment application for experimenting with direct and indirect prompt injection in LLM-assisted CV screening.
Hands-on lab exploring pipeline security and privilege escalation vectors in CI/CD environments.
Deliberately vulnerable AWS cloud infrastructure for practicing cloud security exploitation.
Send us a message with the resource name and link. We review every suggestion.