Expose vulnerabilities in real-world AI systems with new CTF Pack

The pack is designed to provide essential training that prepares teams to identify vulnerabilities before adversaries exploit them in production environments.

Each challenge in this pack represents a real vulnerability class from OWASP LLM Top 10, OWASP ML Top 10, and cutting-edge AI security research, requiring participants to exploit everything from access control systems and financial authorization to ML classifiers and federated learning networks.

The first seven challenges focus on practical LLM exploitation using prompt injection, agent manipulation, and MCP server attacks - techniques accessible to traditional penetration testers. The final four challenges introduce ML model attacks including adversarial examples, gradient leakage, federated learning backdoors, and LoRA artifact exploitation for those ready to advance.

Set up your next CTF!