MultivariateStatistics, Ellipsoids, and other delights

Question

Using the slowly-assimilated-by-the-borg MultivariateStatistics package, one can compute confidence ellipses using the EllipsoidQuantile function. This function returns an Ellipsoid, as you'd hope it would, but, it appears that MultivariateStatistics goes through the trouble of redefining Ellipsoid on its own.

To wit:

Then:

So it gets 'enhanced' (but not shadowed) with this fancy new version as supplied by MultivariateStatistics. Thats nice, but I sure would like to use the regular old version of Ellipsoid. The current version sort of only works ---

(note 1. red shadow syntax highlighting for MultinormalDistribution and 2. the ellipsoid isn't styled (filled) ala the WL built-in one.) but if I want to change styling of the Ellipsoid the down-valued version complains about stuff in the charting package ---

Any ideas of what to do here, best-practice-wise? I can think of a few things involving unprotecting, clobbering, etc, that might work... but was wondering if there was a 'right' way.