I thought about this more… @Clave and other signers should only be publishing the kind:24133 RPC event, not the signed event itself. Have you seen a signer in the wild publishing as well? You're basically talking about a signer offloading the signing AND publishing responsibilities from the client. Some clients might want this — e.g., if the signer manages the user's NIP-65 outbox, the client wouldn't have to worry about relay management. Would essentially be a new sign_and_publish RPC method. Worth discussing and/or experimenting during the nip46 hackathon.