Swati Khandelwal / The Hacker News : Researchers are paid $5K for reporting a bug to OpenAI that would let attackers run code in logged-in sessions of all who opened the hacker's ChatGPT share link — A malicious ChatGPT share link could run an attacker's code in the logged-in session of anyone who opened it …
Researchers are paid $5K for reporting a bug to OpenAI that would let attackers run code in logged-in sessions of all who opened the hacker's ChatGPT share link (Swati Khandelwal/The Hacker News)
In short
Security researchers discovered that malicious actors could manipulate ChatGPT share links to execute unauthorized code. Victims only needed to be logged in and open the shared URL for the code to run in their session. OpenAI awarded the reporting researchers a $5,000 bounty for identifying the critical security issue.
Key points
- Attackers were able to execute code directly inside active ChatGPT user sessions.: Attackers were able to execute code directly inside active ChatGPT user sessions.
- Compromised share links served as the primary mechanism for triggering this exploit.: Compromised share links served as the primary mechanism for triggering this exploit.
- The disclosure was handled through OpenAI's established bug bounty security program.: The disclosure was handled through OpenAI's established bug bounty security program.