You can drive a poseable mesh off of one of the grippable sphere / box / cylinder components, ends up cleaner as otherwise you are going to be looking for overlap with bones on grip and gripping specific bones on the same mesh. Using these invisible components not only lets you define specific collision sizes for the grip, but also run additional logic easier.
Seperate meshes aren’t required, just pulling a % of travel off of the range it is allowed and driving a bone or animation off of it works fine.
Its also difficult to make something generic enough with defining grips on specific bones which may be chained so that it works for all uses, it is far easier way.
Anything super specific like following a spline can be done with a CustomGrip type anyway (though I think I will be adding a spline follower component soon).
Edit Or were you talking about making a custom grippable actor with a poseable mesh base so that it is easier? Because that is a valid point if so since they are only components in engine by default.