Chris,
Try not to introduce a single point of failure while trying to solve for a very unlikely event.
I prefer totally separate circuits for each set of coils and each set of injectors, plus fast and easy access to the switch.
KISS (no, that's not a signature line)
John