I think that's a web search i really don't want on my browser history....

Seriously tho good to know.

simply I'd say yes.

Pets take statinfos etc. from the caster, so pen should be taken from the caster aswell (like sharp staff stat)

But simply go and test it, slot a sharp staff let the scamp hit the dummy -> check max DMG hit
then repeat with another trait like defending -> check max DMG hit
They should differ due to the Penetration difference

I tested some stuff.

I was wearing Mother's Sorrow on body, now monster set, no jewelry and just weapons without set bonusses. I just let the pets hit the dummy and I tested things like Defending Staff vs Sharpened Staff or at first I didn't even skill concentration (light armor passive)...

Sharpened seems to increase the pet damage.
So does concentration from light armor.
So does the Lover mundus stone.
So does Major Breach.

I didn't calculate the exact increase to see if the damage was scaling properly, I just checked if there is an increase in damage by adding one bonus at a time.

There is just one odd thing, sometimes I get inconsistencies. This might be due to minor vulnerability and because I only looked at the avg and max hits on cmx (I was lazy).

So, I've come to the conclusion that pet damage is fully affected by resistances vs spell penetration. If you want to be sure by calculating the numbers and making a more proper (time consuming) testing environment I suggest to do it yourself or hope for someone of the well-known theorycrafters to have a look. But I think you don't have to "over-test" everything.