Over geologic time you can circularise the orbit of a decaying moon by tidal friction, but this would require Phobos and Diemos to be differently circularised. [because tidal forces vary as 1/r^3, and complex tidal torques as even higher powers!]

Alternatively, on a no-plate planet, the Tharsis bulge, as it grew, could have caused the crust to shift wrt the main body of Mars, or for the spin axis to migrate.

There are several asymmetric impact craters, some exceedingly elongate, that attest to bodies in orbits that do not match the modern equator. Unfortunately, they do not tell a consistent story, but there do appear to be more low-angle oblique impacts preserved on Mars than on the Moon.