ChatGPT share links enable code execution read private chats

Researchers at Parameter discovered a vulnerability in ChatGPT where malicious share links could execute attacker code within a logged-in user's session, allowing access to private conversations. OpenAI received the report on October 18 and patched the flaw eight days later, rating it P1 (most critical). Users should avoid clicking ChatGPT links from untrusted sources until confirmed current.