From what I understand atm, I think in general it does make sense to implement some of this logic in a backend canister.
I think there might be some open questions wrt what you would want to check in the validator function vs later, but I think this is now also discussed in the other forum thread?
Also, note that some of the governance methods, e.g., calling manage_neuron for some neuron, can only be called by principals that have permissions on the neuron (we call this the neuron’s “controller” or “hot key” in the NNS, in the SNS this is a bit more flexible). So depending on what you need here, I am not sure if this can be done by a canister that does not have priviledges on some of the neurons.
As mentioned I am currently not sure how this discussion overlaps with this one that seems to be a step further. So I suggest we continue on the newer thread except if there is anything missing here that is independent of the new discussion. Does that make sense?