Interesting, I'm not sure if a computer has the ability to recognize something as a "natural move but also a blunder." It would require a very human-like way of thinking about moves, which computers don't generally have.
Probably the easiest case of "natural move but blunder" is anything that is a top 3 engine move when looking 3-5 moves deep, but losing significantly on deeper evaluation.
Also, this sort of categorization is at the heart of how chess puzzle collections are automatically assembled. A good chess puzzle contains an unnatural move that wins--the exact opposite of the natural but blunder. Chess sites scan their online play databases for these all the time, and serve them up as puzzles.
Any human 1500+ can recognize the natural human move. The way the computer thinks about moves is different.
I really don't believe that Stockfish can tell you "this move is tempting, but wrong."
I'm sure you could build something in that might kind of work, but until I see it I'm skeptical.
Blunders are exactly what a device like the one described would seriously help with. If the buzz means both "there is an only move here and it's not immediately obvious" and "at least one of the natural moves here is a blunder or very inaccurate" then you need to just send a buzz and you've probably cut inaccuracies significantly. That said, a very simple communication device like this is probably badly hurt by a 15 minute delay.