the fighting system would be a really hard thing to make fun.
you would need to introduce RPG elements that improve or augment your characters ability to judge their enemies.
maybe you could implement a feature like the spying they do in the last episode. you could try to stealth around looking at what your opponents do in a battle and figure out a pattern.
maybe ask around town if people battled against them and how they got beat or how they beat them.
the issue here being, once you fought them and lose to them, you would basically know what to do when you reload/retry again.
so you would either make it a rogue-lite where you have to fight your way to the top of a randomly generated set of characters with their own patterns, or you would need to make it a canon thing that each time they fight they change their strategy and you would need to relearn their patterns over again, each time you lose against someone.
box mods could also either be like in the show where your character has a defective box and you can't use them yourself, or you could make them like an item you can get in order to make certain "boss" fights easier... maybe have a sidequest that gives you the item once completed.
so I could see a possibility of making this into an interesting topdown Pokemon like RPG/Adventure game, but it would need other stuff besides just 1 on 1 battles in order to actually be interesting to play imo.
you could also make it a Point 'n Click adventure, with the box peek battles being bossfights only, and in order to know what to do or when to peak you need to aquire information you get by solving puzzles around town (maybe you need to get some item an NPC wants and for bringing it to him you would get information etc.), with the box peek battles acting as stop-gates.
the Insult Sword Fighting from Monkey Island comes to mind, where you needed to find the perfect comebacks, and in order to have them you needed to learn them first and then use them at the right time.
just as an example for a fight, you sit in your box (btw. a point n click could also use the same scenes the show uses when they sit inside the box thinking what to do) and you heard that your opponent is scared of cats because when he was 5 a cat peed on him and he smelled bad for the rest of the day, earlier in the game you solved puzzles and did stuff for some NPC characters to get enough Marbles to buy a "Cat Urine" smell jar (I mean, there are Gorilla vomit ones... so why not cat urine? lol)
so you go into your inventory, open the jar and wait for 3 seconds (maybe another item you require to even do this is a gas mask or something so you don't get sick from the smell yourself???), then you pop out to see your opponent trying to run away because of the smell...
that's just something that I made up real quick, so not very imaginative, but I think you get the point
a point n click could also use basically a 1 to 1 rendition of the shows artstyle, you wouldn't need to make 3D models or adapt the style into a topdown perspective since point n clicks already usually have semi static scenes like the show.
