Here’s a script I’d love to write one day (but sadly no time for the foreseeable future).
First define some rendaku rules, e.g. these https://en.m.wikipedia.org/wiki/Rendaku
Then, for each vocab, check if it follows the rules. If not, add a warning to the wanikani page for the vocab.
Even better if we can think of a consistent mnemonic (eg a little devil) for bad rendaku and add it to the mnemonic.
A good start would be to write a python script that checks kanji against the rules and fills out a spreadsheet for reference at least.