Questions re: Cursor Server (Remote SSH)

[privacy]

hey guys, so, I was watching the cursor-server that is running on a SSH remote hosts. AFAIK this cursor-server is making sure the processes you launch keep running in command line, the state of the IDE is persistent even after reconnects, and so on.

Looking at the resources this cursor-server is allocating, there seems to be a lot going. And that makes me wonder, what information about the cursor user is stored on the server? Cursor userid, email address, oauth, ssh pub key?

See, not all remote hosts that you connect to can be trusted, and imo the user should have some more control about this, like an option to remove the cursor server data when I want to, easily without searching all across the filesystem. Or even better, the option to redact or anonymize personal information about the session.

I’m freelancing a lot - and I just want to make sure that when I handover the project, I can remove my personal information from that remote server entirely with nothing left.

Could you give me some detailed info about this?
I would highly appreciate it!

[danperks]

The cursor-server is a kind of mini-Cursor binary that manages the connection between your local Cursor instance and the server you are SSHing into.

High resource usage is not unexpected, really, as many of the extensions you might have installed have to run on the machine where the code lives, and the cursor-server Binary is the one that would run these - so if you have a lot of extensions installed, that binary will use a lot of resources.

There is actually a command to trigger Cursor to uninstall the cursor-server from the machine you have SSH’d into, which will clean up all the extensions and anything it has cached specifically for this case!