I'd love to see some modification to the auto-mask process for handling faint background stars. I've seen the tips in the documentation which are helpful.
The idea would be if there was some way to "sample" an existing background star that would sort of be a guide for the automatic process to better identify what a faint background star, for lack of a better word, "looks" like. The idea would be to strengthen some of those stars a bit in terms of luminosity.