Are standard errors under MLR in two-level models computed using the approach in Section 3 of Yuan and Bentler (2002). ON NORMAL THEORY BASED INFERENCE FOR MULTILEVEL MODELS WITH DISTRIBUTIONAL VIOLATIONS. Psychometrika. If not, would you provide a reference to the method used by Mplus.
White, Halbert (1980), "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity", Econometrica 48 (4): 817–838,
These standard errors are known as the Huber–White standard errors or robust standard errors because they are more accurate under model and distributional misspecification. This robust approach can be applied to any model (multilevel or multivariate etc.)