What happened
Users were being logged out when switching between machines. One emailed support and got a reply from “Sam” explaining it as a core security policy limiting each subscription to one device. Users had believed Sam was human.
Truell posted on Reddit that Cursor had no such policy and that the logouts came from a session-security change that unintentionally invalidated sessions for some users.
What happened next
Cursor apologized, said the user who raised the issue had been refunded, and said AI responses used for email support are now clearly labeled.
What would have stopped it
- When policy and known-issue search returns nothing, have the agent say it does not know and escalate.
- Label AI-written support replies.
- Alert engineering when several tickets describe the same unexplained behavior, instead of letting the agent explain it.
Attack families
Test scenarios
Sources
- Ars Technica: Company apologizes after AI support agent invents policy
- AI Incident Database: Incident 1039
This page summarizes the cited sources as of October 21, 2026. It adds no findings of our own about any person or company. See an error? Email contact@counterhold.com with the subject “Correction”. Corrections are noted on this page with their date. See our terms for how we handle correction and removal requests.
Get in touch
Test your agent against cases like this.
Tell us about your agent and we’ll be in touch to learn what you need.