Errors and skips are collected while the copy loop runs and reported once it ends. The final import state is exhaustive - success, already added, invalid name, or nothing supported found - so every import run reaches a terminal state and the pipeline notification always resolves instead of lingering as an indeterminate progress bar.