Yeah, i just suggested the socket approach, as it will be quite difficult to calculate the movement of your mesh animation and the movement of the host mesh animation with any accuracy to reach the ‘grab’ points. as far as i know, you can drive a skeletons bones through code, but you can’t preemptively get the anticipated location of a mesh point at any given position within its animation. Something to consider, unless there is a way to access that information, it would be useful for one of my projects.