I’m going to completely agree with you on this one, @nckmnn. This is one of my biggest complaints about Wanikani, and I say this as someone who’s “completed” it and hit level 60. I’m not bashing here, this is an area in which I’d like to see some improvement.
The way it should work, and presents itself as working, is using the component parts (what Wanikani calls radicals) to create memorable mnemonics. Like you pointed out, the simplest way to break down 暗 is Sun + Sound. But for some reason, Wanikani breaks it down even further. And “Sound” is a radical that exists in the system!
In this case, it might just be an oversight, but there are a lot of examples from later levels where Wanikani reuses a combination of radicals when they should really be combined into one. Take 襟, Collar, which is so obviously ネ, Spirit, plus 禁, Prohibit, but Wanikani teaches as Spirit + Tree + Tree + Jackhammer. That’s just dumb. (Edit: Oops, actually it’s 衤, which should really also be its own radical. My bad.)
Luckily, there is a solution! The single most useful script out there, in my opinion: The Phonetic-Semantic Composition Script. Not only does this let you see the actual component parts that went into creating the kanji (for example, 滋, 磁, and 慈, which all share the root 茲 not taught on Wanikani at all), it also makes learning the readings a hundred times easier (滋, 磁, and 慈 all share the On’yomi じ, for example).
This is all information that really, really needs to be included in the native, vanilla Wanikani experience. And again, I say this as someone who clearly approves of Wanikani enough to “complete” it. In the meantime, @nckmnn, @M4tthi4s91, and anyone else who feels the same: install the Phonetic-Semantic Composition script and vastly improve your kanji-learning experience instantly.