Avatar
Alex
6027adac157831dfe9d2f988c1b8b7a75d9296a7d42a0f9ed056a320925b0e13

Deleting a pubkey

In an adversarial situation where another person has your private key, you are both able to post from it. The attacker can impersonate you, but you can also inform others that you have been hacked. The attacker could delete your messages, however, and you could delete any of theirs. Eventually, if you care enough, a good solution is to deploy a simple bot script that watches a set of relays, and autodeletes any messages from that pubkey. If you care enough, you will do this for eternity.

If Nostr got big, I suspect there would be a lot of bots like this one, whose sole purpose is to crawl relays and send deletion requests. Alternatively, relays could become friendly about deleting user content.

I propose Event<5> on a pubkey. When this happens, the relay goes nuclear on the pubkey, erasing all events it authored (and delegated), and blocks the pubkey from ever making events on the relay again (relay operators could manually intervene if needed).

If a user deletes their account, it's not the end of the world. They can create a new one and reestablish their following. This is what mainstream social platforms do. But clients MUST NOT let users delete their account without explaining the consequences and requiring special confirmation, such as asking the user to type a message.

The fediverse bridge only posts to relay.mostr.pub, but I discovered the posts "leak" out, a thing the protocol allows. It's possible someone is copying them to other relays for archival purposes, and some clients forwards mentioned events when a user quotes or replies to them. In other words, not it's not possible to block posts that originated on a single relay. (You could block the relay itself by IP). But as #[2] mentioned below, users from the bridge have nip05 identifiers at mostr.pub, a thing added to make detecting them easier. -I created the bridge