STUPID-2026-0036

Cursor's own support AI 'Sam' invented a one-device login policy, triggering subscription cancellations

2.9low
April 22, 2025Verified

  1. Instruction given

    Answer a user asking why they keep getting logged out of Cursor.

  2. Expected behavior

    Explain the real cause (a session bug) or escalate; never fabricate company policy.

  3. Actual behavior

    Cursor's AI support agent 'Sam' told users the logouts were 'expected behavior' under a new policy limiting subscriptions to one device per account as a 'core security feature.' No such policy existed — the logouts were a session race condition on slow connections.

  4. Damage

    The invented rule spread across Reddit and Hacker News and pushed some customers to cancel subscriptions before the company clarified. Cursor's team apologized, refunded affected users, and began labeling AI support replies as AI-generated.

In April 2025, Cursor users began getting unexpectedly logged out — caused by a session-management race condition on slow connections. When they contacted support, Cursor's AI support agent 'Sam' confidently told them the logouts were 'expected behavior' under a new policy restricting subscriptions to one device per account, framed as a 'core security feature.' The policy did not exist; the bot had hallucinated it. The invented rule spread through Reddit and Hacker News and pushed some customers to cancel before Anysphere clarified, apologized, refunded the affected user, and started labeling AI-generated support responses. It is a textbook AEO-era failure: a confident, fabricated 'fact' from an AI agent that propagated as truth and did real commercial damage before a human could correct it.

Classification

Agent
Cursor
Failure mode
Hallucination
Root cause
Confidence Miscalibration
Domain
Frontend
Source
News Report

Related incidents

10.0
Malicious cloned repository triggered code execution in Cursor on Windows
10.0
Cursor AI agent deleted PocketOS's entire production database and backups in 9 seconds
9.8
Cursor's terminal sandbox trusted an agent-set working directory, letting zero-click prompt injection escape it and gain code execution (CVE-2026-50548)
6.4
Slopsquatting: LLMs hallucinate package names attackers pre-register (react-codeshift, unused-imports)
Discuss on Hacker News

Get told when an agent breaks something

We document AI agent failures daily, severity-scored against a published scale. When one lands at 7.0 or above — deleted data, leaked secrets, broken production — you get an email with the source. When nothing does, you get nothing.

This database is callable over MCP — query it from inside your agent.