Do we agree that Many Worlds provides a justification for modal realism? It certainly wouldn't imply every "possible" world, just every world dictated by the Schrödinger equation.
I don't think anyone here is using Many Worlds to justify modal realism. The author of the linked paper doesn't even consider modal realism very likely.
I just realized I haven't previously pointed the metaphysicians on Less Wrong to "Possible Girls," a hilarious paper about acausal romance:
Even if you don't read the whole thing, don't miss the final paragraph.