Alright, I spent probably longer on this then I should have, but good news!
Short version: it basically works the way I thought it did.
Long version: there are some things to make sure you set up correctly, and I observed a roughly +/- 2 bpm rounding error when transitioning between bpms that I’ll explain.
I ran a MIDI clock from my computer (let’s call it SOURCE) into an MC6 Pro (this should work the same as an MC8 Pro, which wasn’t easily accessible at the time). Make sure that the MIDI Thru is set so that the MIDI Clock from the source can go through the controller to the final device (let’s call it ENDPOINT), e.g. I had the USB Device route to DIN, which then went to another controller I was using to monitor the MIDI clock output).
As long as the SOURCE produces MIDI clock, the MC6 passes it forward. When it stops, the MC6 stops sending clock, even with MIDI Clock Persist on, BUT (1) the ENDPOINT already has the tempo, so in most devices, it just keeps that and (2) the MC6 does store the final bpm, so if you tell it to start its MIDI Clock, it will start with that stored bpm. So, basically, if you can stop the SOURCE via MIDI, you can set up a switch that stops the SOURCE and starts the MC6 clock, and it should be seamless.
Once the MC6 Clock begins, you can set it via Tap or via preconfigured BPM values. If the external clock SOURCE comes back, the MC6 clock stops again and the MC6 passes forward that according to its MIDI thru settings.
Now about that bpm wobble: whenever MIDI clock stops sending, the bpm at the attached device tends to be based on whatever it’s most recent average of BPM values is. So, for example, if SOURCE is sending 125bpm then stops, if either end was a bit behind at the moment it stops, then you might see the ENDPOINT with ~123bpm until you turn on the MC6 and it again sends continuous clock data.
Hope that helps.