Basically muscle memory is the same as explaining the lift on air plane wing with the Bernoulli theorem.
It is a nice shortcut but really has not that much to do with what actually happens
Muscle are not equipped to remember anything and do not have anything that can activate motor neurone directly.
Even reflexes like "tendons reflexes" a receptor have to send a message to the spine in order for the motor neurons to be activated.
That is as far as muscle "memory" goes
It is certain that repetitive exercise has an influence on which and when motor units are fired but that depends on motor neurons not muscular fibres.
As well motor leaning have lasting effect on neuron of the motor cortex which dictates what motor neurons are activated in what order. it seems that "memory cells " has been recently put in evidence by Dr. Emilio Bizzi and his collaborators.
(Li C-SR, Padoa-Schioppa C, Bizzi E: Neuronal correlates of motor performance and motor learning in the primary motor cortex of monkeys adapting to an external force field. Neuron 2001, 30:593-607.)
If you are interested there was an interesting thread on the flinch where reflexes vs cognition was mentioned.
The muscle memory you are referring to is literal muscle memory.
You are absolutely correct that this does not exist as such. I think, however, most people refer to "muscle memory" as the process of acquiring learned motor pattern programs.
I did not realize that there was a separate "muscle memory" that referred to body building/strength training. Thank you for pointing that out as well.