JailbreakZoo is a repository focused on breaking large models, including large language models and vision-language models. The project aims to explore vulnerabilities, exploitation methods, and defense mechanisms of these advanced AI models to promote a deeper understanding and awareness of security in large-scale AI systems.