It's a nice concept, but the design makes it so the correct choice is to always pick the mode the card says (at least if the AI chooses randomly). Winning an attack, winning a defense and winning a support should do different things. So when you draw this super good defense card, maybe you still want to try support - because you really need to heal right now! Or when you get this great attack card, you don't attack because the opponent has a shield this turn, or something.
Right now it seems more random than meaningful choices
Viewing post in Plan of Attack jam comments
The theme of the game isn't that much important. Actually. you should go for the theme that fits the mechanics the most. There seems to be something missing from the design in terms of mechanics. I'd say if you want to make this into a complete game, you have to complete these gaps first.
Will the player build his own deck like in MTG? Will he build it and improve it incrementally like in slay the spire? If you improve the game to release it, I'm not even sure I would keep the "only one card" part of it, because even if it is an interesting constraint in the context of this game jam, I don't think it would be a lot of fun in a complete game. It removes choice, and choice is important.
Maybe the cards could each have a good side and a bad side. Like the card has an attack symbol, and says if you win you deal 2 more damage. But if you lose, you take twice as much damage. Now playing it as attack or not is a real choice. Then again I'm not sure how interesting this would be. It still feels more like random guessing than strategy.