Generate a collision exactly the same (Or in x,y,z dimensional format) of the mesh and skeletal mesh?

Is there any way to generate a collision exactly the same (Or in x,y,z dimensional format) of the mesh and skeletal mesh using box trace by channel/Sphere trace by channel, etc…to detect collisions while I position the object similar to " Get component bounds and draw debug line" ?

Like this for example:
Capturar
But detecting collisions