Or is it needed at all? JMdict and EDICT clients have a kanji database available which has all the 常用 status marked. If it is necessary to flag a word as containing non-常用漢字, it could be done on-the-fly.
While I'm here (in my befuddled 3:28am state) one (never going to happen) wishlist item I have is a linking of kana to individual kanji as far as they can be individually linked. At the moment I might suddenly want to find out what words have 御 read as み. It's easy to see what entries have kanji headwords with 御 and it's easy to see what entries have readings with み but (unless 御 is first in the entry) it is non-trivial to go any further.