Agent encountered error

Support Help
1

image
image972×376 18 KB

image
image1196×576 57.1 KB

6

Hey, thanks for the report, and sorry this happened. I saw the screenshot with the error “We encountered an unexpected error repeatedly”.

I checked your agent. This happens when a Cloud Agent started as “New Project” later gets a real project name. Renaming changes the repo URL, but the original agent still points to the old one, so every follow-up in that chat fails with “Agent encountered an error”. We’re tracking this issue, and your report is attached to it.

Good news, your code is safe. Everything the agent managed to do is in the renamed repo under your account. To keep going:

  1. In Cursor, start a new Cloud Agent.
  2. When picking a repo, choose your project under the new name, not “New Project”.
  3. In the first message, briefly say where you left off. The new agent will start from the current state of the repo.

You won’t be able to resume the original chat, but nothing was lost in the repo. I can’t share a timeline for a fix yet, but I’ll post an update here if I have one. Let me know if the new agent gets stuck on anything.

8

我也遇到了遮掩个问题，怎么解决？

9

我也遇到了这样的问题，怎么解决

10

Hey @li_zhiqiang, the message “Agent encountered an error” is generic and can happen for different reasons, so we need a bit more info:

  1. Is this a Cloud Agent? Did you create the project via New Project start from scratch, and then rename it?
  2. Could you share the Request ID or the full error text? Especially if it says something like [not_found] Repository .../tmp-... was not found on Origin. Also share your Cursor version and OS.

If it’s the same case where the agent was started as a New Project and then the project was renamed, your code is still there. It’s in the repo under the new name. To continue:

  1. Start a new Cloud Agent.
  2. In the repo picker, choose the project under its new name, not the tmp-... entry and not the old project in the sidebar.
  3. In the first message, briefly say where you left off. The new agent will continue from the current repo state.

If it turns out to be a different cause, it’s better to open a separate thread with a screenshot and the Request ID so different issues don’t get mixed up. Let me know how it goes.