An adversarial LLM that stress-tests AI systems via exploit prompts, uncovering vulnerabilities like bias, data leaks, and jailbreaks. Designed for ethical AI security research. Open-source, developer-focused, and ethics-driven.