It’s good to know the VESC is robust enough to handle loss of power during operation. I still don’t think we should ever cut power to our motor controller. I equate this to turning off your PC by unplugging it from the wall. Each time you do this you put the VESC at risk not knowing what internal state it is in. Partially through instruction execution could be an irrecoverable state. I propose it is better to use the VESC as a motor controller - when you want the motor off send a signal which turns it off.
If a motor is still spinning for 1-2 seconds under water after signaling off it is not spinning with any dangerous torque. With no additional energy being given to the motor it is spinning based on a finite amount energy (momentum of rotating body) and that finite amount of energy is quickly diminished by friction between the water.
If you set recuperation > 0A you will also have a EMF helping to decrease momentum after a stop as well. This is what @nice2cu is reporting.
I’m proposing you do not cut power. You use a relay to switch the PPM signal input to the receiver. The signal is switched from remote signals to signals from a signal generator like an Arduino. The Arduino generates signals which correspond to the motor not spinning. The relay is wired so it uses these signals when low and the remote when high.
This approach does not rely on any undefined behavior such as leaving the PPM signal input floating