I can observe very slow execution of virtually any terminal command that is executed by LLM inside chat. The problem doesn’t appear if I try manually execute command inside cursor terminal or any other dedicated terminal. All network logs complete in the range of milliseconds.
Steps to Reproduce
There wasn’t any specific change that was applied to introduce this.
Expected Behavior
AI assisted commands should execute at roughly the same time of manual execution.
Screenshots / Screen Recordings
Operating System
MacOS
Current Cursor Version (Menu → About Cursor → Copy)
Can you tell me how can I log information for you? Or you already know what you are fixing? I observed that things start to slow down immediately after I switch the working folder. I fresh installed multiple times Cursor. I was able to run every single time commands (initiated by llm) without any slowness, just as I would manually do this before I switch to any folder. Right after I switch workspace (new chat for llm), commands start to execute slowly. Currently have:
Version: 1.7.17 (Universal)
VSCode Version: 1.99.3
Commit: 34881053400013f38e2354f1479c88c9067039a0
Date: 2025-09-29T03:10:26.099Z
Electron: 34.5.8
Chromium: 132.0.6834.210
Node.js: 20.19.1
V8: 13.2.152.41-electron.0
OS: Darwin arm64 25.0.0