Agents window: automation sessions crashing

[lo_ol]

Where does the bug appear (feature/product)?

Cloud Agent (GitHub, Slack, Web, Linear)

Describe the Bug

prompting a session created by an automation from the agents window causes an error

Steps to Reproduce

see above

Expected Behavior

dont crash

Screenshots / Screen Recordings

Operating System

MacOS

Version Information

Version: 3.11.13
VS Code Extension API: 1.125.0
Commit: 3f21b08f0b436a07be29fbfe00b304fa15553350
Date: 2026-07-10T01:45:28.254Z
Layout: Agent Window
Build Type: Stable
Release Track: Default
Electron: 40.10.3
Chromium: 144.0.7559.236
Node.js: 24.15.0
V8: 14.4.258.32-electron.0
xterm.js: 6.1.0-beta.256
OS: Darwin arm64 25.5.0

Additional Information

Please fix your testing process. I pay money to use this product in a functional state. I am not an employee of cursor and I do not appreciate being treated as free QA on EVERY release. There are constant obvious bugs that clearly show no one bothers to even run some of the code that is getting merged. Please address this

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

[deanrie]

Hey, thanks for the report. This looks like a known class of issues. When you do a follow-up on a session created by automation, the backend recreates the VM and the git branch, and that step can sometimes fail. In the UI, it shows up as the generic “We encountered an unexpected error repeatedly”. We’re tracking it, but we don’t have an ETA for a fix yet.

To confirm you’re hitting this exact signature, please share the link to the failed agent cursor.com/agents/… or its bcId. With that, the on-call can pull the turn-workflow logs.

Workarounds for now:

• Retry the follow-up. Sometimes it’s a one-off and the retry works.
• If that doesn’t help, start a fresh agent instead of continuing the old session.

I know this kind of instability gets in the way. Send the bcId and we’ll take a closer look.