Dry Run

Trusting a scheduled agent by watching what it does over time means it was already doing that unsupervised the whole time you were watching. Something has to check its work before the first run, not after enough of them.