converting vertex animation shape keys to key frames perfectly using pseudo rig

In an effort to archive a historic model appropriately, I’m attempting to export it in the Collada DAE format with vertex animation stored in sets of key frames made with shape keys.

However, I could not export the key frames because the Collada format only supports skeletal animation key frames.

I found a script that creates an armature with deformation bones for every vertex (1872 total).

How can I automatically set the positions of each bone to match the position of each vertex in a shape key?

Are there any flaws in my plan, and if so can they be circumvented?

I’m willing to upload an example if needed.