If you assume the device coming on means the button has been pressed, that means you cannot control that device in any other way. That is that device would be limited to always being part of a specific scene.
That’s pretty antithetical to the SmartThings philosophy, which says that rules are not stored in the device definition, but multiple rules can be applied to the same device under different conditions.
I’m not saying don’t do it, I’m just saying I don’t think there would be much demand for it, it’s not what people are generally looking for from SmartThings. But if you find it useful for yourself, or just is an interesting project, you can do it.
Note, however, that SmartThings cannot load the scenes into this controller, for all the reasons discussed above. You’re just going to have to manually set up the scene in the controller and separately write whatever code you want for SmartThings, maybe using core, to react to the various zwave devices coming on.
(But you aren’t going to do this with the device type handler for this device, because this device doesn’t report its events to the hub. Which in some ways makes it easier. You can just ignore the fact that the Cooper device exists when you write your logic to react to the other and devices coming on. The Cooper is basically going to look like a manual control for those devices from the SmartThings point of view.)
Honestly, though, unless you already own a bunch of these, it’s going to be easier to pick any of the button controllers off of the list of the ones that already work well with SmartThings. 