Sometimes commands just spin and I cannot tell if they are running or not. They’ll go on for minutes and at a certain point I just cancel it. When I retry is usually, but not always goes through.
I get it, it’s frustrating when it’s unclear if something is frozen or just taking a long time. This is a known UI issue with long-running commands, and the team is aware.
To rule out other causes (network lag, indexing issues):
Can you check:
Cursor Settings → Network → Run Diagnostics, and share the results
Cursor Settings → Indexing & Docs, make sure indexing isn’t stuck in a loop
And we need a bit more info:
Which exact commands are hanging (npm, git, build commands)?
A screenshot of the stuck state
Output panel (View → Output), any errors in different channels?
Thanks for the quick response. So my theory is given other things working simultaneously and it working immediately as soon as I tried in a new agent window then it went right through then it’s unlikely it was network or index. But obvi IDK.
It was an aws mcp command. I wasn’t paying too close of attention. I think it was filtering a cloud watch log.
Unfortunately the screen is long gone. I was hoping the request ID would be good enough.
Do you mean the agent output or the terminal output? Either way it’s lost now, but I can grab it next time.
