How did stockfish not see mate in one??

As you can see from the analysis bar on the left, Stockfish does not see a mate in one for black with white to move. Even if you do Rxf7+ Kxf7, there is still absolutely nothing that white can do to win.

Edit five minutes after writing this: I acknowledge that 1. h3 Rxf1+ 2. Kh2 exists, but then the continuation is 2. ... Rdd1 3. White pushes a pawn, and black wins with 3. ... Rh1. White can make as many escape holes with their pawns as they want, but it's always a mate in a few moves.

Yes, it is the 3500+ Elo chess engine that is wrong. The engine is always going to be right. Just follow the moves the engine tells you and you'll see why it isn't mate in a few because I guarantee if it was anything less than mate in 10, the engine will spot it nearly instantly, even on default settings because it goes to depth 20, which is 10 moves.

How about white plays g3? Stockfish will show the line if you're open to it.

You can play this position against the engine if you select 'Practice vs Computer', but I'm going to warn you it won't be easy. The position is around -7/-8. (depth 40-ish), which is absolutely winning, but you'll need all the advantage you can get when you are a 500 rated player against the computer. They're the best defending players in the world. And white has three connected passed pawns on the queenside. I'm not even sure I would bet on you winning this position.

Stockfish, a highly advanced chess engine known for its powerful analytical capabilities, sometimes encounters situations where it misses a straightforward "mate in one" due to a variety of factors. These can include the complexity of the board state, the depth of its search algorithms, or specific settings and limitations of the engine at the time.

Unless the position is illegal, stockfish will not miss mate in one 99.999999999999999% of the time and it will never miss the mate after it is played (attempt to play on after mate), so you could always test it any time you think it is missing mate in one. The only way an engine misses any mate in 10 or less is if it is a composed position that purposefully attempts to manipulate the algorithms of the engine to "trick" it into missing mate.

Of course, if you put limitations on the engine, it could miss anything depending on the limitations. This is a normal position in a real game, with standard settings. Not a relevant issue.

This. It’s still quite a long way from actual checkmate and it would take you many moves to beat the fish from that position, if you could even do it at all.

However the broader lesson from this for the future is it’s (almost) never “why did stockfish not see” etc etc and instead “stockfish evaluation isn’t what I expected. what am I not seeing?”

If the engine was always right ches would be solved now

There is no mate in 1 for black (though black has a strong advantage). As you say, it's white's move and they should play g3. The analysis view should show this.

What ended up happening in this game?

White played Rc7 and got mated.

Wait what? How did you know that? That's exactly what happened. Did you look through my recent games or something?

: U didn't need to cut me off

I have an old version of Stockfish running on my old laptop. In the original position, he sees mate-in 7 in less than 1 second. Your mate-in-1 doesn't exist.

By the way, your Rdd1 move fails after the obvious fxe3, and black doesn't have "mate in a few moves" any more. Still up a rook though.

General advice: if you think that Stockfish is wrong and you are right, you should think again.

You might have input the position wrong because there is no mate in 7, nor is there a piece on e3 that can be captured.