Why would you write a plugin at a point where you can't choose to disable the kepub features when sending a book to the device? And it sounds like you're going to be reimplementing a good chunk of some of calibre's conversion pipeline. I already need to decide if I want to process CSS on my own, and I'm leaning toward "no way, let calibre handle it". Code should be re-used whenever possible, but only copied when it's essential.
It strikes me (as a purely personal opinion, perhaps wrong because text fails to convey intonation and emotion) that you're being a bit rude to davidfor. He's trying to help you make your plugin better. It is a good thing to play Devil's Advocate to make sure issues with whatever is currently known get brought up and thought about. It also is an excellent way to make a good developer consider whether they are approaching the problem in the right way. It doesn't mean you won't get help, just that you need to think. Go look at the .kepub.epub thread in the Kobo reader forum, davidfor disagreed strongly with the database code, but he gave me some really good information on how to proceed with making it better anyway. He also disagrees with me having the kepub processing enabled by default, but I agree with his reasoning on keeping it off by default; the only reason I don't is because it's the entire point of my plugin to do that processing. Maybe you feel that you're getting criticism instead? Good developers take polite criticism and use it to improve the code. It's not always easy, but it's for the best. You were even given a link that gives you a good chunk of the code you're going to need.
TL;DR - I'm not convinced (and I may not be alone) that a file format plugin is the way to approach this, but if you want to do it anyway you'll find lots of help if you want it.