MUAs and MTAs should talk more

I will not go into great detail, here, though perhaps requests in comments would get me to say more.  For now, though, the short version:

I have come to hold firmly the opinion that Mail Transfer Agents (MTAs) and Mail User Agents (MUAs) should talk more.

For example, it should be possible for me, from my MUA (currently mutt, though I'm thinking of switching to something emacs-based) to flag a particular message as spam in a way that permanently blocks all messages to the recipient that that mail was destined for.  Now, obviously (I hope) this only is relevant in cases where one is using a wildcard accept type of rule, i.e. where any username at the domainname in question will get delivered to one particular actual user's mailbox (with perhaps some exceptions going to different users, or, as implied here, being rejected as spam).

Now, perhaps folks just aren't doing this very much any more, and I should get over the idea that all MTAs everywhere should have this capability.

For what it's worth, I wish at least Gmail for domains would do it...  and have something procmail-like, too...  or rather, something more elaborate than the current filtering rules, which I find limiting.  Though, perhaps I'll give it a go, we'll see.

(I'm currently in the market for a new mail infrastructure.)