I've included my own thoughts in the annotations.

2.Nc3 (?) against me is punished by 2...d5 where the c-pawn break is cruelly missed to get any kind of advantage as White.

6..c5 (?) seem dubious. The strenght of the same line in a regular KID is that the bishop on g7 can't be blocked by a pawn after 7.d5, but here where c2 has not moved there is always c3, killing the fianchettoed bishop.

11...c4 (?) seems to create a weakness. I would play ...Rd8 12.Be2 (forced, isn't it ?) Bf5, with a fine game.

14.Qc2 ? calls for ...Bf5 : Qe1 is better (forced ?).

14...b5 (?) (your sideline) is not an improvement compared to the game (15...Bf5 !). First, 15.b3 is not forced (15.b4 is stronger). And even after 15.b3, Black's structure is more damaged than White's, and the queen reaches a safe square.

16...b4 ? gets a weak pawn on c4 for nothing, when the simple 16...Qxa3 gives a (real) structural advantage and a good vs. bad bishop after 17.bxa3 Bd3.

25.Rxc4 ? falls for the fork, while bxc4 keeps a good pawn. And 26.Rc7 ?? decides not even to fight with Rfc1 (forced, Rcc1 is too sad) with pawn vs. exchange and microscopic chances to draw due to the prisoner on g8 ; but the rooks are really active.

Thanks for the analysis, I learned quite a bit.