I would still recommend FFT as the standard choice. You typically only need to modify the Poisson solver if you introduce special dielectric or metallic regions.

We don't have COOP analysis yet, it's on the idea list for the future though.

Integrated DOS can be computed easily just using the raw DOS data, although I guess the details depend a bit on how you define the integration, like is it only over energy, or you want it k-point resolved, etc? Note that you can also compute molecular eigenstates (for finite systems) and Bloch states (for periodic cases) with QuantumATK, this gives a direct 3D picture of the eigenstates.