What about a mechanism analogous to a CrO3 oxidation? This generates a Cr(II) species which can then reduce a Cr(IV) species (electron transfer reaction?), which generates two Cr(III) species. See attached scheme. If you want to generate 2Cr(III) straight off the bat you will need to proceed via radicals as any fully ionic mechanism (ie. full curly arrows rather than fish-hook type) will always do redox in multiples of two
I would have thought it's a reasonable suggestion. If you consider the generally accepted mechanism for a Cr(VI) oxidation, a Cr(IV) species is initially generated which presumably then disproportionates - since it is well known that CrO3 oxidations ultimately leave you with Cr(III), which is why they go green.
What do you think?