I don't think intergrating UVW Randomizer or Coronatriplanar maps into the physical material would be the best way to go, as they are not always used by all materials and could weigh the material down with unnecessary code (one argument could be that other properties are also not used by all materials, like vol. absorption, clearcoat or sheen but those are more confined in their properties).
But I 100% agree that there has to be a better way to batch edit bitmaps, maybe with the source transform input you mentioned. I would add that ColorCorrection is something I often have to use (to mass make the saturation or hue of leaves for example in several bitmaps of a MultiMaterial), and it gets quite crowded and unintuitive very quickly, and if you want to add a Multimap to those maps it becomes quite the chore.
I think a logical first step would be to add those randomization parameters to the Coronabitmap itself, and then to enable mass editing of similar maps, where one would select 5 Coronabitmaps or Colorcorrection maps and would be able to adjust the shared properties on the fly (except some specific ones, like bitmap source etc). I don't know if that's possible or a limitation of the Slate Material Editor.
An alternate tidier solution would be to have multiple inputs and outputs to said maps, where one could plug 5 Coronabitmaps to one Coronauvwrandomizer or Colorcorrection, and could draw 5 outputs that correspond to the inputs, (similar to the way Multimap handles inputs).