Checks whether value held across an issuance topology stays fully backed IN AGGREGATE, not merely per account, as balances move between the buffers you declare. You name each buffer (its role, its asset type, and what it backs), declare a backing-ratio requirement and per-buffer floor/ceiling, and declare a set of movements. This tool evaluates the aggregate backing verdict before and after those movements, flags any buffer that breaches its floor or ceiling, reports the thinnest safe margin per buffer, weighs the declared idle-balance cost against the declared crossing cost, and identifies the specific movement that first breaks the invariant, if any. Settlement-asset agnostic: the same computation runs unchanged whether the settlement asset is centrally issued, pooled-account-backed, or reserve-backed, because every fact is a caller-declared input. Not every settlement asset has a buffer set to check: for a direct or two-tier CBDC the holder's claim IS central bank money, so there is no backing question. You declare this with the backing model selector; a declared vacuous model returns BACKING_NOT_APPLICABLE, a defined answer distinct from both an intact pass and a shortfall.
The demo figures below are synthetic and already loaded, so you can press Run without entering anything of your own. Load the composition-break demo to see the flag this tool exists to catch: a movement that leaves the buffer set's total balance unchanged but still breaks the backing invariant.
It does not sweep or optimize a treasury — 106 already ships a Sweep Optimizer across five accounts with ACH/FedNow/SEPA-Instant/CHIPS cut-off times over a 13-week horizon. It does not net — ART-259 and ART-368 already do that. It does not attest reserves — ART-06, ART-512, and ART-280 supply those facts where the settlement asset is reserve-backed. It issues no recommendation to move money and no financial advice, and it proves nothing about whether the declared balances match any external ledger.
A movement can leave a buffer set's grand total balance completely unchanged and still break the backing invariant, because moving value out of a buffer that backs circulation into a buffer that does not changes the COMPOSITION of the backing, not just its location. That composition shift is invisible to any per-account check and is the one thing neither the sweep workbench nor the netting nodes above look for.