I stop trusting a clean distribution table the moment a finalized row can still look open after it was already consumed once.

A delegate executes the claim. That should end the row. Later, ops sees that same row sitting in the batch queue, finance freezes the release, and support now has to prove whether this is a valid settlement or a second payout for a right that was already spent.

That is the SIGN problem that felt real to me. In TokenTable, the row is versioned and immutable once finalized, so the argument is no longer about which export is current. The argument is about history. Did the delegated execution already consume the entitlement before batch settlement picked it up again?

That is the bug I care about. Not the failed payment. The payout that still looks legitimate because two execution paths both leave enough room to say yes.

When money is blocked, row history has to prove one exact thing: the earlier delegated path already spent the entitlement before the batch path queued it again. If it cannot prove that, the second "valid" payout gets expensive fast. #SignDigitalSovereignInfra $SIGN @SignOfficial