The dual transistor is configured as a comparator, sensing the voltage difference between its emitters. It will have a linear transition between low and high on [U14:4] over a small range centred on a point where V[U14:5] is fractionally lower than V[U14:6]. When V[U14:5]>>V[U14:6], [U14:4] is pulled up to near v[U14:1], turning the MOSFET off. When V[U14:5]<<V[U14:6], R3 pulls the MOSFET gate down, turning it on.
Max current is set by the MOSFET's Id_max and dissipation due to I2R losses in its Rds_on, though if there is insufficient gate drive voltage, losses in its body diode may take over.
Max voltage has four constraints - the ideal diode cathode voltage is limited by U14 Vcbo_max and Q3 Vgs_max, and the maximum reverse voltage across the diode is limited by U14 reverse Vbe breakdown and Q3 Vds_max. None may be exceeded.
I would not expect it to do well at eliminating transient over-voltage on its supply when feeding motor drivers as it is unlikely to react to spikes fast enough to make a difference. It certainly wont stop the supply dipping and browning out the MCU during sudden acceleration of the motor.