basilisk

Basilisk is an open-source AI red teaming framework designed for automated adversarial prompt testing against various large language models (LLMs) such as Claude and GPT-family models. It features evolutionary prompt search, structured attack modules, and a real-time scan dashboard, enabling security researchers and penetration testers to conduct repeatable and comprehensive security assessments of LLM applications. Notable capabilities include differential mode comparisons across multiple model providers, guardrail posture scanning, and the ability to export test results in various formats.