Hey, thanks for the report. I see the screenshot, this is a known issue where the agent can’t match the file context with the patch and gets stuck retrying.
Related thread: Agent stuck in editing loop
The team is aware, but we don’t have a specific ETA yet.
For debugging, can you share the Request ID from the chat where this happened? Top right of the chat > menu > Copy Request ID. This will help us understand why loop detection isn’t triggering.
As a workaround, try switching the model from Auto to a specific one and see if it behaves better.