I’ve definitely been bitten by this. You always have to weigh the chance you break something against the upside. If you’re actually fixing a bug, fine. But just refactoring to make something cleaner? Or deleting because it seems like it’s not doing anything, even after doing some research? Think again.