It's just a
python script, which loops through things in particular categories.
I only have a small number of cards which are considered "staples", which go in every deck (by default), and those are largely personal preference. So for White it would be
Tithe, Tutor,
Sentinel. Of course I sometimes take them out. I have a bit of logic for multicoloured, like put Emeria only in monowhite decks.
The rest, it does have a ranking in that it just goes down the list. I haven't thought very hard about the order of the list. For example, right now at least the removal list goes:
So in a monowhite deck you'd get Swords, then whatever's way down below for monowhite later. In WB you'd get Anguished and
Utter End too, if it was BG you'd only get
Assassin's Trophy, etc. In practice the order doesn't actually matter that much at the very top anyway because I'm going to add like eight removal spells or whatever so who cares whether Swords or
Path to Exile is top? Both will end up in monowhite decks. It's only around the cut-off point. And anyway, like I said, it's a template, I tend to switch stuff in and out.
Dispatch for example is not on this list but becomes really good when metalcraft is likely.
It's not sophisticated enough to make smart decisions about rounding the suite out given the colours or anything, so multicoloured literally just amounts to "add what you can, in this order".
The removal list probably isn't long enough because it usually gets down to
Introduction to Annihilation in monocoloured which I don't really like. But then, it only has a few monocoloured removal spells on there, particularly for some colours. It's primarily creature-based removal, too, it's not sophisticated to say "make sure you have three artifact removal as well" or whatever (could be, I guess, I'd just add a category or something)... it gets hard though when some colours just don't have good options for some things. But then, I'm probably going to have to actually think about such things in that situation anyway, and this is just to get a baseline.
Frankly, the big win IMO is in doing the lands.
Ain't nobody got enough time to keep adding all the different duals and stuff each time.