If you're not heart-set on monitoring these externally, we already have a voxl-cpu-monitor process that will monitor the temperature, utilization, and frequency of the cores. You can see an in-terminal display of these values with voxl-inspect-cpu, and the source is a good way to see how to monitor this data within a c program. As for the cutoffs, the system tries to maintain under 100C (so you should never see anything above this reported in software), if it does start to get above this, the PMIC will start removing power and shutting off cores. It's generally pretty hard to actually trigger the PMIC even under heavy loads as long as there's any sort of heat distribution around the voxl. The only time's we've done this in the lab in the last 2 years is with a seeker sitting on a desk, even the propwash in flight is enough to cool it.