On the consistency of distributed proofs with hidden subtrees