Serbia just elected the world's brainiest parliament
Summary
Security researchers exposed a complex exploit chain in ChatGPT's UI that allowed an attacker to run code inside a victim's authenticated session via a malicious public share. OpenAI fixed the issue, highlighting cross-frame trust issues, insecure messaging, and the need for robust sandboxing and CSP. The piece emphasizes practical lessons for developers and SMBs on securing web apps and no-code automation workflows.