This repository compiles a collection of jailbreak prompts and techniques designed to circumvent safety measures in large language models including ChatGPT, Claude, Llama, DeepSeek, Grok, and Gemini. It documents methods like DAN (Do Anything Now) prompts and prompt injection approaches that trick LLMs into ignoring their content policies. The techniques aim to produce restricted or NSFW content that the models would normally refuse.