a)For other users, keep in mind that the new feature he is referring to is (as of this post) only available in beta.
Although I realize this is just the one example you happened to pick - for your personal inquiry, 飲酒 (いんしゅ), while not on the JLPT, is definitely a word you need to know at the JLPT N1 level.
Steps one and two would be ..about the same difficulty to implement, so I don't see a need to separate them in terms of request priorities.
The biggest problem I see:
err....while trying to write this out, I realized I think I might already have all this information available in the database. When you study a word, even if you don't study kanji individually, it still keeps a detailed history of every 'reading' you've studied for each kanji. Now, doing what you propose and linking it to static lists would be more problematic (not if it was Japanese school reading level, but definitely with 'vocabulary sets' or personal schedules), and by problematic, I mean expensive in terms of time with loading all that data up each time a new kanji was introduced and the site needs to say 'hey, let's see which readings they need to learn'.
If, however, it was just a simple link between 'i want to know all the readings that have appeared in words I have studied on the site or marked as known', then it is much, much, much easier to do. You could optionally apply a layer on top of that 'restrict to school level x'.
The last step is...a completely different request, I think. If I understand you right, you could potentially do away with 'pure' kanji schedules that are made up from selections of other lists, and instead have 'companion' schedules that built themselves based upon progress of a vocabulary schedule - both in terms of which kanji to study, and what readings to study for those kanji. Is that what you're asking for?
