Looking at the provided installed unit, I see it being connected via some small cables to the distribution unit. As the incoming feeders are 10mm at least, so rated for 100A or so, and the blue box is only connected by what looks like 2.5mm cables at best, i would hazard to say it can only correct for a current of around 20A at most, otherwise the wires would burn out pretty fast.
Hard things needed are maximum harmonic current ( can be true RMS or even just equivalent current, does not matter so long as the measurement method is given, you can work the maths to get a better idea given the details, but a Fourier transform would be a lot better as well) the unit can inject or remove per cycle, and the maximum instantaneous ( either over a single mains cycle or even per sampling period if period is defined) current it can handle.
Now, as to overvoltage spec, the typical lightning pulse is given an IEC spec, which is where you see that 8uS pulse rise time on EMC testing, given that lightning typically hits high voltage ( 132kV or higher) power lines, and is attenuated by both line impedance and by a lot of lightning arrestors. Those are arc gaps on each HV pylon ( those little horns across each insulator hanging the line), the massive VDR stacks at each distribution point ( looks like a very fat insulator and contains thousands of varistor capsules stacked on each other), the distribution transformers themselves ( makes a low pass filter to broaden the leading edge and attenuate the pulse and spread the energy in time for the VDR stacks to react), and then further arc gaps ( smaller gap so lower breakdown voltage), more VDR stacks ( also smaller, as they only have to withstand 66kV, 33kV or 11kV operating voltage) and also line reactors to filter harmonics and of course the smaller transformers at point of use points. Still, with all that, you will get kV pulses on your incoming mains, and just how high a voltage can this clamp, given that high voltage power devices trade off power handling for speed, reliability and rate of rise, and miller capacitance is going to give you a bad day every time you exceed bonding wire limitations.
Now, if you can give a box that handles 100A, even single phase, can handle harmonic currents up to 10kHz, and can be connected via a 100A breaker and not fail when you put on a 50A vibratory feeder ( the thing most hated for power factor correction, as it uses a half wave power SCR to drive a massive electromagnet, giving you every harmonic but the fundamental mains frequency and spikes as well with the phase angle control deciding where on each pulse to turn on said hockey puck SCR with 20A of gate current to turn it on hard and fast) on the power line, along with handling the power factor correction for the 22kW motors driving the conveyors on the line, and you will sell these like hot cakes.