"why Black's position is not just comfortable but actually winning"
++ White made a mistake before. 6 h3? is loss of tempo and weakening. 6 Bg5 is fine.
8...Be7 is still fine: black can play 9...Be6.
"after 8...Be6, if White plays 9.Bg5, Stockfish is suggesting 9...Qa5, and it seems to consider 10.Bxf6 a major mistake, inferior to 10.Bd2 and indeed even inferior to 10.Kf1. I'm surprised that Black can so happily allow his pawn structure to be wrecked."
++ 10 Bxf6? cedes the bishop's pair and opens the g-file for black.
In a recent rapid game I was playing Black against the Bowdler Attack in the Sicilian Defense. The game proceeded as follows:
Here I played 8...Be7, which is all right, but Stockfish prefers 8...Be6. In fact, letting Stockfish play against itself at classical time controls, it seems that the White position cannot be saved after 8...Be6. The next few moves in Stockfish vs. Stockfish following 8...Be6 were 9.Nf3 h6 10.O-O Bd6 11.Re1 O-O. I suppose 10...Bd6 in this variation explains why my move 8...Be7 was inferior. But I have trouble understanding why Black's position is not just comfortable but actually winning, in principle, after 8...Be6.
A further oddity: after 8...Be6, if White plays 9.Bg5, Stockfish is suggesting 9...Qa5, and it seems to consider 10.Bxf6 a major mistake, inferior to 10.Bd2 and indeed even inferior to 10.Kf1. I'm surprised that Black can so happily allow his pawn structure to be wrecked.
I would be most interested to hear any explanation or insights into these lines.