I think that you are correct that the dark areas are areas that the crack existed prior to the failure and the light areas are the areas that failed when the blades separated. I don't think that the dark color is oil soak, it's just a different type of crack surface.

My new theory: I did not sand the flashing off of these props. I think that the sharp edge of the prop, combined with the stress riser of the crease, combined with loads due to holding a blade while wrenching on the prop-nut, caused a crack in the blade at the crease.

This crack grew during operation and then failed.

(I know, technically composites don't fatigue. It's just a theory.)