A leak in ChatGPT Lockdown Mode let you forward your chat history, OpenAI has patched it
Check Point Research found a way to bypass Lockdown Mode, the optional mode in ChatGPT designed to prevent data leaking through prompt injection. The flaw was in the HTML preview of code blocks: that sandbox allowed pop-up windows that fell outside Lockdown Mode, so an attacker could use a manipulated prompt, GPT or shared chat to siphon off your chat history. OpenAI reported on 18 September 2026 that the flaw had been fixed; Check Point published the details on 5 November.
What does this mean? Users don't need to do anything, but it shows that even an extra security mode is no guarantee: don't enter confidential client or patient data into a consumer account.







