I can’t offer a perfect way to do this, but you may be helped by looking at this another way.
The photometry aperture containing the light of the supernova will itself contain background light from the host galaxy. That disc containing the light of the supernova will be huge in terms of the galaxy, covering many millions if not billions of stars. Hence the trick is not to subtract a sky background without the galaxy, but to have an annulus that is representative of the background light you are collecting with the supernova photons.
Andy