The poster you originally replied to is suggesting this for revoking the attackers keys. Not for revocation of their own keys…

there's still some risk of publishing an attacker's key. For example, what if the attacker's key had access to sensitive user data?