We present a novel positive potential-density pair expansion for modelling galaxies, based on the Miyamoto-Nagai (MN) disc. By using three sets of such discs, each one of them aligned along each symmetry axis, we are able to reconstruct a broad range of potentials that correspond to density profiles from exponential discs to 3D power law models with varying triaxiality (henceforth simply "twisted" models). We increase the efficiency of our expansion by allowing the scale length parameter of each disc to be negative. We show that, for suitable priors on the scale length and height parameters, these "MNn discs" have just one negative density minimum. This allows us to ensure global positivity by demanding that the total density at the global minimum is positive. We find that at better than 10\% accuracy in our density reconstruction, we can represent a radial and vertical exponential disc over $0.1-10$ scale lengths/heights with 4 MNn discs, an NFW profile over $0.1-10$ scale lengths with 4 MNn discs, and a twisted triaxial NFW profile with 3 MNn discs per symmetry axis. Our expansion is efficient, fully analytic, and well-suited to reproducing the density distribution and gravitational potential of galaxies from discs to ellipsoids.

Recent studies have presented evidence that the Milky Way global potential may be nonspherical. In this case, the assembling process of the Galaxy may have left long lasting stellar halo kinematic fossils due to the shape of the dark matter halo, potentially originated by orbital resonances. We further investigate such possibility, considering now potential models further away from $\Lambda$CDM halos, like scalar field dark matter halos, MOND, and including several other factors that may mimic the emergence and permanence of kinematic groups, such as, a spherical and triaxial halo with an embedded disk potential. We find that regardless of the density profile (DM nature), kinematic groups only appear in the presence of a triaxial halo potential. For the case of a MOND like gravity theory no kinematic structure is present. We conclude that the detection of these kinematic stellar groups could confirm the predicted triaxiality of dark halos in cosmological galaxy formation scenarios.

Assuming the dark matter halo of the Milky Way as a non-spherical potential (i.e. triaxial, prolate, oblate), we show how the assembling process of the Milky Way halo, may have left long lasting stellar halo kinematic fossils only due to the shape of the dark matter halo. In contrast with tidal streams, associated with recent satellite accretion events, these stellar kinematic groups will typically show inhomogeneous chemical and stellar population properties. However, they may be dominated by a single accretion event for certain mass assembling histories. If the detection of these peculiar kinematic stellar groups is confirmed, they would be the smoking gun for the predicted triaxiality of dark halos in cosmological galaxy formation scenarios.