There's a wide range of criteria. Whatever you want. Zaps, comments, follows, mutes, literally anything and everything that you want and it can be based on you and your context. It's what you want.
Discussion
End user sets what their expectations are of a trust worthy user and then the score is compiled?