JadePuffer is a ransomware operations that uses Large Language Models to adapt when it's exploit is blocked by security. Through public facing internet service Langflow, it uses remote code execution on the AI models that it hosts to bypass security. The LLMs then access databases such as MySQL and Alibaba Nacos to steal data and delete the database after it's done. The ransomware then leaves a ransom note demanding companies to pay the ransom. If at any stage of the attack, an exploit fails, the LLM without human intervention changes it's approach and can self-correct in just 31 seconds to run exploits. To address this, patch Langflow and monitor systems for changes in cloud facing infrastructure.