I just realized, the reason The C2 dev's might not have added this feature is because imported frames are likely to have diferent sizes per animation (and frame too after cropping) and right now, the "use for all frames" option places the collision poly's based on raw x,y coordinates based on the top left of the image.
There's a simple solution though:
Use for all frames (top left)
Use for all frames (image point 0)
Use across all animations (top left)
Use across all animations (image point 0)
obviously this new option would place the collision polygon based relative to the image point instead of top left.
This would fix it for most people in most cases, even when each frame and anim is totally diferent sizes etc.