Bug Description:
When using the “Auto” agent mode, the AI gets stuck in an infinite loop. It keeps repeating the same message: “I will search for relevant test files,” without making progress. It seems to hallucinate and never actually analyzes the code.
Steps to Reproduce:
- Open a test file
- Enable the “Auto” agent mode
- Ask the agent to analyze or refactor the test
- The agent keeps repeating the same reasoning loop and doesn’t proceed
Screenshot:
OS & Cursor Version:
- OS: windows 11
- Cursor version: 0.49.6
Does it block usage?
Yes — the loop makes the Auto agent unusable for test analysis or refactoring.
request ID : 01220256-41da-4437-88c9-a71495f50263