This is a very solid and interesting manuscript, which addresses the proper modelling of the transition between brittle and ductile failure. Publication is recommended, but the authors may wish to pay attention to the following issues in a revised version:

* While viscosity (mostly) remedies mesh dependence, it should be pointed out that it is a weak regularisation technique in quasi-static loading conditions, as considered here. Only under dynamic loading conditions full regularisation of the governing equations can be proven, see De Borst and Duretz (2020). This should be discussed briefly in the paper in order to avoid misconceptions in the community.

* Work on combining mode-I and mode-II plasticity has been pursued before in the context of modelling of concrete, albeit for plane stress rather than for plane strain or 3D conditions as is the focus of the current paper, see e.g. Feenstra and de Borst. Int. J. Solids Structures (1996) 33, 707 - 730. It would be advisable to put the present contribution also in that context.

* The discussion on volumetric locking is confusing and perhaps misleading. It is suggested that volumetric locking occurs for isochoric plastic deformations and it is not (explicitly) pointed out that this phenomenon also occurs for dilatant or contractant plastic flow. Indeed, in all cases a kinematic constraint is imposed at failure, i.e. when the elastic deformations vanish, see De Borst and Groen, Int. J. Num. Meth. Engng (1995) 38, 2887 - 2906. 

This is a robust manuscript looking at the addition of a tensile cap for modelling rock failure in a geodynamic context. The manuscript is of excellent quality and should eventually be published to benefit the community. There is however a point where the manuscript could be improved, regarding viscoplasticity and strain regularization.

- The manuscript presents Perzyna's viscoplasticity as a way to regularize the ill-posed problem of strain localization. While this is valid way to consider viscoplasticity, there is no mention in the manuscript that viscoplasticity could also capture a physical behaviour of rate-dependent plastic deformation (which is not uncommon).

- The authors claim that introducing viscoplasticity in a quasi-static approximation is sufficient to regularize the strain localization problem. I am not aware of any studies proving this statement. A number of studies analyzed the role of viscoplasticity for strain regularization and, to my knowledge, they all included the dynamic terms.

- Furthermore, the authors should consider discussing other studies about viscoplasticity and strain regularization, some of which are more critical about the role of viscoplasticity than the cited studies in the manuscript (Jacquey et al. (2021) and Stathas and Stefanou (2022) as examples).

References :

Stathas, A. and Stefanou, I. The role of viscous regularization in dynamical problems, strain localization and mesh dependency. Computer Methods in applied Mechanics and Engineering, 388, 113185, 2022.

Jacquey, A. B.,  Rattez, H. and Veveakis, M. Strain localization regularization and patterns formation in rate-dependent plastic materials with multiphysics coupling. Journal of the Mechanics and Physics of Solids, 152, 104422, 2021.