As cited
Copy frozen at (site build).
vulnerabilities
A Shared ChatGPT Link Could Run Attacker Code in Your Session and Read Private Chats
Researchers at Parameter discovered a vulnerability in ChatGPT's share link feature that allowed malicious links to execute attacker code within a victim's authenticated session, potentially exposing private conversations. OpenAI received the report on October 18, patched the flaw eight days later, and assigned it the highest priority rating.
Why it matters: Any ChatGPT user who clicked a malicious share link risked unauthorized access to their chat history and sessions; practitioners should review whether they clicked unfamiliar ChatGPT links and consider reviewing their chat history for sensitive information shared during the vulnerability window.
- Source published
- First seen by Cybersecurity Tracker