Collection metadata commands have different error reporting behavior in some cases (when run multiple times, for example). We should standardize when an error is reported back to the user, and audit the current set of commands to make sure they meet this standard.