@changil The spine-unity runtime uses standard Unity components like a MeshRenderer to render a normal (procedurally generated) Unity Mesh. Spine covers only animation and mesh generation, collision detection and raycast intersection is still done via Unity's components.
You could e.g. create a GameObject with a BoneFollower component and let it follow e.g. your upper-arm bone, then attach a BoxCollider or BoxCollider2D component as you would with normal Unity objects. Then collision and raycasts can interact with the box collider as ususal.
If you want to use drag-and-drop to remove a cloth-item from the character skeleton to undress it and drag it e.g. into an inventory slot, I would recommend against moving any bone around. You would normally generate a new independet GameObject for the dragged cloth item. This could be either a Unity Sprite (a static image) or a new SkeletonRenderer GameObject with just the respective cloth-item skin active.
