AnalysisCybersecurityOctober 31, 2026

Shared ChatGPT link could run attacker code in victim's session

Read original source →thehackernews.com

Parameter researchers showed a malicious share link executing code in a logged-in ChatGPT session, listing recent conversations and exfiltrating a private chat. Reported via OpenAI's bug bounty on October 18, fixed eight days later and rated P1, its most serious category.

1 source

More stories today

Open the live feed