Woof, so instead of some standard way to store agent vs human blame (like in Git metadata), this is all only stored on Cursors servers ..

> When you view files and commits, data is fetched from Cursor's servers.