Thanks for the comment
Despite the similarities with Aurora there are huge differences under the hood.
Now Aurora already had an experimental CTF mode but I added a lot of improvements to the gameplay, the most visible change is the fog of war which lead to other fundamental changes.
Basically the AI is entirely new, It understands the goal of the game better than in Aurora but it doesn't cheat, the AI doesn't have knowledge the position of the players pieces but it does count probability from all the information the game gives and it uses the starting position as a reference and keeps updating, so for example if the ai temporarily sees one your pieces such as the bishop but then the queen moves out of vision the AI tries to narrow down the possibility where the piece might be. It trades if it thinks it's a good trade or a free piece but ultimately focuses on capturing the flag and recapturing its own if it's taken.
There are also other aspects that are related to both the AI and fog of war that are not instantly obvious for example if a piece moves throught the opponents line of sight (their pieces threat lines and move squares) that piece is temporarily visible during the movement animation, for human players the squares that were tracked will play the movement animation few times in a loop, AI also has this information, basically piece "a" moved from "b" direction towards direction "c" that means that this piece then must be in one of the squares of "xyz".
That was huge amount of work and now the AI is pretty good, still needs some work, mostly how to get the captured flag to it's own base
I basically created a custom 3d chess engine out of which I build Aurora and Raid Flag from