ChatGPT Lockdown Mode bypass let attackers silently steal private chat histories
VULNERABILITYChatGPTredteam/daily · 2 sources
- Severity
- High
- Sources
- 2
What happened
Researcher Alexey Bukhteyev published findings on 5 November 2026 showing a bypass of OpenAI's Lockdown Mode in ChatGPT. The flaw let attackers silently exfiltrate full chat histories, deliver malware and phish credentials from within the trusted interface. In the ChatGPT Classic desktop app, the leak was triggered by normal conversation with no further user interaction.
For IT people · Written by AI from 2 sources · Wrong?
What to do
- Check whether your ChatGPT Classic desktop app has received a fix from OpenAI, as the article does not state one.
- Avoid sharing sensitive information in ChatGPT conversations until the fix is confirmed.
Written by AI from the sources, checked against them
Full story timeline · 2 updates
Who reported · 2 outlets
Cyber Security NewsGBHackers
Who and what is in it
Companies