I used AI coding tools to create a single-player logic game called Darkenn. It is like a single player version of Go.
Every move flips one cell and its neighbours. Under the hood it is linear algebra over GF(2), and the solver computes a minimum-move solution rather than just any solution.
I have been personally drawn towards games and what they teach us. Darkenn teaches us seeing dependencies over one action and its flow towards the final goal.