"I've never seen this thing, except in this obvious and common case where it happens all the time, but other than that, literally never".
Mate :/
In general, my first step in deciding whether projects are good is to google "<project name> review", and see what other developers and sysadmins have experienced with it.
I have often found that a 5 stars system contains inadequate information density to provide me with guidance on the suitability of one project over another with respect to integrating it into my companies tech stack, although I appreciate that's a personal standpoint and that you may disagree.
A project with lots of stars and rare commits is almost never better than one with a less stars but low-grade continuous activity.
"Date of last commit" is probably FAR higher than any "social" bullshit in terms of quality.
That's the first thing I look at when I arrive on a github page, always.