Feature request for product/service
Chat
Describe the request
I’ve been running automated tests through Cursor’s browser tools where agents follow detailed specifications in markdown files.
These tests often involve repetitive but necessary actions (like voting sequences in a matching algorithm test), and the agent loop detection keeps stopping the execution prematurely.
The issue is that the system flags legitimate test sequences as “looping” when they’re just following the test specification. For example, having 8 players vote through a specific matrix results in similar UI interactions that get flagged.
Could you consider adding an option to disable agent loop detection?
A simple toggle in agent settings or in the chat would solve this.
Thanks for considering!