Collusion within and across helper party networks

[eriktaubeneck]

I think that this probably needs a little text on collusion within a set of aggregators vs. within a set of coordinators vs. across both aggregators and coordinators. That is, we don't consider the role of a given helper party to affect this requirement, though some protocols could be more robust to compromise of a greater subset of the whole.

Originally posted by @martinthomson in #14 (comment)

[eriktaubeneck]

Probably also between helper parties and first/delegated parties.