The Great Consensus Cleanup proposes to make 64 byte transactions consensus invalid, for various reasons. They have been non-standard for years and appear to serve no useful purpose.
It appears that any such transaction can be malleated into a different size, so that in practice there is no confiscation risk even if such a transaction were useful. I’m trying to understand this aspect better.
Given a UTXO A, which either already exists or is freshly mined under the current consensus rules (e.g. through a non-standard transaction).
Given a 64 byte transaction T that spends it (non-standard by definition).
Without access to the secrets (if any) that were used to generate T (private key, hash pre-image, etc.)
Is it always possible to change T into a consensus valid transaction T' that spends A?
Additionally, if I understand correctly, it’s not possible for any 64 byte transaction T to send coins to a destination B that is a secure (i.e. >= 20 byte) public key (hash).
Regardless, for every B, is there a malleated T' that preserves B (at least still spendable by the private key b)?
If the latter is too hard to prove, could one could show that any T can be malleated by an attacker to divert coins away from B.
