Most of the time the command run by the agent in the inbuilt terminal just sits and does nothing. Only way to proceed further is pop out the terminal. How to fix that? Is it possible to always run command in 1 attached terminal? so the agent can follow all the terminal progress?
I am running 0.46.11 and with claude-3.7.sonnet.
TIA
bump. Did you figure this out? I don’t know if you’re describing the same issue, but I’m on Windows 11, just started using cursor last month so I’m fairly noobish atm, although I’ve been a professional software dev for 12+ years now. Issue I’m having is that the agent will run some terminal commands in an embedded agent terminal it creates for itself and then try to run some other commands in a terminal tab in the IDE env outside of the chat window. This causes issues because it’ll change to the project root in the embedded window and then switch over to the IDE tab to execute the command, but the directory hasn’t changed to project root in the IDE tab terminal. It’s driving me nuts.
