Hi David,
I would not bother attempting flat correction of slitless spectra. Normal flat correction techniques do not work because each point is a mix of spectral and spatial information. (The technique used on professional setups is complex and involves generating effectively a “flat cube” for each wavelength at each point in the field.) My advice would be to take a conventional flat beforehand to check for any obvious dust donuts etc and place your target and reference star zero orders as near as possible on the same pixel avoiding any dust spots so any gross flat effects eg vignetting etc cancel.
Cheers
Robin