Of course I am sure.
Read what the main developer of Duniter said (he said it about ucoin, which is the previous name of duniter)
https://forum.duniter.org/t/objection-in-the-model-of-the-web-of-trust-of-ucoin/764/2uCoin is already compliant with what you propose, because it can be reduced as a pubkey selection problem which is a pure human decision and is not relevant at uCoin software level.
Because uCoin have no idea what a human is, this is an undefined term. uCoin only recognize public keys and relations between them. The WoT rules are here as requirements to avoid massive cheating, and will work even if pubkeys are selected during assemblies. The assemblies would just be an help for people to have confidence in the existence of a unique human behind the key (which is a thing that remains to be proven - I am not convinced).
But If you really want vertical control over the pubkeys that can be members of the community, then you will have to make another software, or at least fork it to behave this way.
Just my opinion.
So you can take the WoT code of duniter (which is currently applied for humans), and apply it to assemblies.
Each assembly should have a unique multisig public key of course, that should be signed by all the participants of the assembly.
Spit the cancer of censorship and of spamming out of the dash forums.
Let's LIBERATE the users/readers of the dash forums, NOW!