Here's what it does while you're asleep.
It starts by walking the network and writing down every machine that answers, and what each one is running. Then it checks that list against the public vulnerability database.
This is where most scanners stop and hand you a few hundred rows. NK Node keeps going. It tests each finding to see whether it's real. The ones that hold up get reported. The ones it couldn't settle are flagged for a person. The false alarms are thrown out, and you never see them.
Along the way it notices which machines everything else leans on, by watching who talks to whom. Those go to the top of the list.
By morning there's a report on the box. It's short, it's written for a person, and it says what to fix first and why.