We present DSKAG-IT-SIG, a family of information-theoretic transaction signature schemes that achieve unconditional existential unforgeability under adaptive chosen-message attack by computationally unbounded adversaries. The construction derives per-transaction MAC keys through DSKAG, a deterministic symmetric key agreement protocol requiring no key transmission, no handshake, and no public key infrastructure. We prove (Theorem 1) that the forgery advantage of any unbounded adversary making q queries is at most q * 2^-128 in standard mode under the ideal cipher model, reducing to the statistical uniformity of DSKAG-derived keys and the random-function behavior of HMAC-SHA256 under a uniform key. In the standard model, the bound holds up to the PRF distinguishing advantage of HMAC-SHA256. We prove (Theorem 2) that cross-domain forgery advantage is at most 2^-128 + epsilonᵢso, reducing to the statistical key isolation of DSKAG across policy domains. Both bounds are unconditional in the ideal cipher model and independent of any computational hardness assumption. Standard-mode signatures are 30 bytes, a 97. 8% reduction versus Falcon-512 (666 bytes) and compatible with ISO 20022 SWIFT message fields without re-engineering. The NexusKey composite policy digest binds asset class, jurisdiction, KYC level, and chain identity into the key derivation path; policy bypass is cryptographically equivalent to key forgery. A four-layer UltraHonk zero-knowledge proof system (143, 802 gates, no trusted setup, 16 KB proof) enables permissionless on-chain compliance verification, deployed on Ethereum Sepolia and Arbitrum Sepolia. Version 2. 3. 18 pages, 8 tables. Three independent academic institutions validated the construction: no structural attacks found. Changes from v2. 2: concrete finite bounds replacing generic negl (lambda) in Properties 1 and 2; buffer uniqueness derived from txₛeq monotonicity (no SHA3 collision resistance dependency) ; explicit ideal cipher model and standard model dual framing for HMAC analysis; formal separation of empirical and theoretical claims.
Richard Blech (Fri,) studied this question.