Draws

I meant if one of the players had one pawn and gets boxed cause of it and can get checkmated by the other person with the knight

In FIDE, it's a win for the player with a knight vs pawn if the player with the pawn runs out of time. Doesn't matter which pawn, it just has to promote to rook, bishop or knight to allow a help-mate. On chess sites, where it's timeout, a knight for the player with time is only draw, since a knight or bishop alone can't force checkmate. Although a help-mate is still possible if the opponent has pieces (except if it's only queens), chess sites don't count both sides' material and therefore N vs ~ with black timeout is a draw.

Yes, white could deliberately flag here to avoid getting mated and get a draw:

......... from chess.com. Not because it's right or fair or the outcome of an intelligent evaluation - but only because the website's programmers team could not handle all such situations. However since a few years there is a pretty good computer program to determine whether a position is dead (no mate is possible not even a helpmate) or alive (mate or helpmate is possible). Once chess.com has integrated this tool in it's operation we can be assured they will change the rule in accordance with FIDE- (chess games) and WFCC (chess problems) rules.

Note that FIDE-laws already demand that chess websites do implement automatic draws for dead positions in online games (article 5.4) so a multi-million dollar fine for chess.com is to be expected

For the 4 man endgame we don't need a special software. The tablebases are more than sufficient.