Agent stuck or not? How to tell the difference?

Every Cursor user has experienced Cursor’s Agent hanging, i.e. stuck.

The challenge is knowing when it’s properly stuck (will never complete), or when it’s probably going to complete if you left it a big longer.

Those 3 bouncing dots are often a good sign that Agent might complete the action, but not always.

Lots of time can be wasted waiting for Agent to complete an action, when it fact it might never (until/unless interrupted).

Anything that Cursor developers can do to help us better differentiate between a “properly” hung command, and one that has a genuine chance of completing, would be greatly appreciated.

(Of course even better would be that Cursor gets much better at NOT stuck in the first place!)

1 Like

hi @spaceman and thank you for your post.

As this seems to be an issue could you post a full separate Bug Report with more info so we can investigate and fix it Create Bug Report

Hey, could you let me know in which situations the agent gets stuck?

  • Does it happen at the beginning of the response?

  • When it calls a tool?

  • When it executes terminal commands?

  • Or is it something else?

Additionally, please let me know:

  • Which model you are using

  • Your operating system

  • Your Cursor version

can I send you chat logs?

Hey, yes you can, but please also answer the questions above.

Let me document this the next time it happens

1 Like