The problem I want to solve is that if my nsec is compromised, I want a way to flag old messages as untrustworthy and not have to start an entirely fresh profile.
The idea is that seed phrases are used to generate people's nsec/npubs. If their nsec is compromised, the same seed phrase can be used to produce a new nsec/npub pair that is verifiably associated with the compromised profile based on the associated xpub. The xpub would be made public.
Clients would need to additionally verify that a profile's npub is generated by its xpub, if one is supplied.
To recover from a compromised nsec:
I use my seed phrase to generate a new nsec/npub.
I then create a new profile with the new npub, including the xpub.
Clients will periodically search for profiles with the same xpub as any of the profiles they follow. If they find a new profile with the same xpub, they verify that the new npub supersedes the previous, and then can automatically update their follow list, flag all events from the old npub as untrustworthy, and whatever else.
The problem I want to solve is that if my nsec is compromised, I want a way to flag old messages as untrustworthy and not have to start an entirely fresh profile.
The idea is that seed phrases are used to generate people's nsec/npubs. If their nsec is compromised, the same seed phrase can be used to produce a new nsec/npub pair that is verifiably associated with the compromised profile based on the associated xpub. The xpub would be made public.
Clients would need to additionally verify that a profile's npub is generated by its xpub, if one is supplied.
To recover from a compromised nsec: