Idea 1: I’m no expert on the outbox/inbox model (is there a good summary somewhere?); I suppose you’d want to use a client that’s “mute friendly,” meaning that if you mute someone, your client doesn’t post your content to a relay that doesn’t support the muting feature.

Idea 2: I think the idea would be that it’s basically a private account. All rando accounts are blocked by default. Only whitelisted accounts can see content (whether some or all of it). The question would be whether you want to do all the whitelisting yourself. If not, web of trust could potentially help here, by vetting users according to whatever criteria you deem appropriate and whitelisting them (subject to your override of course).