Oxford’s Dynamic Collections is a forthcoming WordPress plugin that seems to still be in private beta, but which sounds interesting. Basically, it harvests OER [Open Educational Resource] records into WordPress from across a range or sources, but filters them by keyword(s). The results, presumably with a bit of hand tweaking, are quickly-built subject lists of such resources.

My guess would be that one could probably do something similar with repository record feeds: use Excel to sort simple CSV records by the presence of keyword(s), then export only the relevant records as CSV, then load these into Omeka.