Researchers detail takeover of ChatGPT accounts
The gist
Researchers at Parameter found a public ChatGPT share could run attacker code in a victim's signed-in session.
OpenAI fixed it October 26 and rated the report P1, its top severity tier.
How it unfolded
- October 2026Parameter researchers found that an ordinary public ChatGPT share could be turned into attacker-controlled JavaScript running inside the signed-in session of whoever opened it.
- Oct 26, 2026OpenAI fixed the issue and classified the report as P1.
- Oct 28, 2026Mukul Goyal's article detailing the takeover technique was published.