Mihail Marin simply missed that 1.Ng6 is winning, even faster than 1.Qg6. During opening analysis, and even during a game, sometimes when you find a satisfying continuation you forget to follow Lasker's advice and look for an even stronger one.
This is a slight mistake in his analysis, but his evaluation of the variation remains correct (White is winning, even faster than he thought) and I see no reason not to continue studying his book and learning from it.
No chess book is completely void of analysis mistakes. If this is the worst one in the book, then the work is still very trustworthy. However, I encourage you to keep your good habit of checking the author analysis, with your own mind and with engine: First, to prevent falling for a bigger hole in a flawed variation if there is one ; Second, to better understand and remember the positions.