In my opinion it's achievable by cheating a little.
We would have to put a virtual wall under the bowl with a coefficient of friction of 1 and also model the content with the right density, then we just assign the gravity and we start the calculation, if the calculation is successful it is that the component is stable, otherwise it is that it tilts.
If you want to push things further, in my opinion you have to go through a dynamic study
I think it should be done with motion simulation by monitoring the orientation of the resulting forces (data that can be retrieved quite easily in motion).
I think that even dynamic simulation does not allow this type of study.
The virtual wall assumes that the object is stationary, sliding (small displacement) is possible if we model the "table".