I'd bet that if you look up the MCU p/n on the web for your controller, you'll find the company making the chip has development info and maybe sample code, etc., which I would guess is what many contrlller makers use as unchanged as possible to save time and money.
But...unless you just cant' find the features you want in an existing controller, *and* you find the chip you're after are capable of that, *and* you really like the hardware in an existing controller for it's power stage, control, etc., for wahtever reasons, you'd probably be better off just finding a controller that meets your needs and using it as-is.
I don't know how hard it is to write code fro those motor controller MCUs; if you ahvet o do it from scratch it might be tougher than doing it froma devkit example; you'll probably have a lot of testing and debugging to do to ensure it will actually do what you want it to (and nothing you don't) in all situations in which you'd actually be using it.
There are existing custom controller projects here on ES and elsewhere, mostly in the motor tech subforum, if you want to see how much they've gone thru to get where they are now (and how many just gave up at some point), and examples on waht to do and what not to do.
