Looks like you got it working, but what I was referring to with the Execute block was adding the full command like this to see if it would override the first.
In the Python tab it looks like the command duplicated. I don't have the AtomQR to test with, so I'm not sure if this works or not...
However, I found another workaround without the Execute block. After loading the AtomicQR example flow, then switch from Atom-Lite to Atom-Matrix in the settings, it will indeed add
rgb.set_screen() on the Python tab. This is the bug. If you just select one of the 25 LED squares in the UI editor to change it to white, then select the same square again to change it back to black. Check the Python tab again and it will now show
If you save that as a new program (ex: Atomic_QR-Matrix), the next time you load the file, it should have the Matrix as the hardware type and the correct rgb.set_screen command in the python tab.