Aurigraph Research Team · Tokenization · Patent-pending (PCT-AUR-CTF) · status as of 2026-10-20
The Composite Token Framework (CTF) is Aurigraph's tokenization primitive for regulated real-world assets. Where a conventional token or NFT records only who owns how much, a CTF token is a patent-pending, five-layer credential that binds decentralized identity, regulatory compliance, provenance, a cryptographic binding to the underlying artefact, and post-quantum key material into a single W3C Verifiable Credential. Every layer is independently verifiable; tampering with any one invalidates the whole envelope; and the credential is designed to be signed with ML-DSA-87 (in the V12 source, AV11-3534; not yet deployed) — NIST Category 5 post-quantum security. Every lifecycle event is designed to be anchored to a Merkle tree whose root is signed with ML-DSA-87 and attested through DMRV (Digital Monitoring, Reporting & Verification); today the registry signing key is generated in memory per tree and is not yet persisted, and consensus anchoring runs in the platform's SIMULATE consensus mode. CTF is the substrate beneath Aurigraph's real-world-asset verticals: hallmarked gold and HUID, CBDC and trade finance, carbon and environmental assets, and tokenized securities. It composes existing W3C and ISO standards rather than inventing new ones — a deliberate choice for enterprise interoperability over lock-in.
SIMULATE
Most blockchain tokens are accounting entries. They prove "this address holds this many units" — and nothing else. That single layer is dangerously thin the moment the underlying asset is something a regulator, auditor, insurer, or counterparty actually cares about: a hallmarked gold bar, a letter of credit, a carbon offset, a battery passport, or a security.
Consider what a plain token — or even a rich-metadata NFT — leaves unanswered for a regulated asset:
The naive fix is to issue separate tokens — one for identity, one for compliance, one for custody — and link them. That introduces three failure modes: links break, links are forgeable, and verification degrades into a slow, trust-fragile graph traversal. A verifier must fetch five artefacts, confirm five signatures, and reason about whether the links between them are authentic.
CTF collapses that graph into one envelope. Verification becomes a single signature check across the whole credential; tampering with any layer invalidates the whole thing; and the verifier returns a single boolean — verified or not — instead of asking the consumer to reassemble five separate proofs.
A CTF token is a single credential composed of five interlocking layers. Each is independently verifiable, and each contributes a distinct guarantee. They are bound together by the outer credential signature (§3) and by the Provenance layer's hash chain, so no layer can be swapped, stripped, or forged without breaking the envelope.
did:aurigraph:asset:<id>
The identity layer gives the asset — and every party to it — a W3C DID Core decentralized identifier (did:aurigraph:...), with its controller, authentication methods, and lifecycle status. Identity is a first-class citizen, not an afterthought inferred from a wallet address. Aurigraph operates its own DID ledger; a DID resolves to a composite token in its own right, whose facets (Identity, Key, Provenance, Compliance) mirror the CTF structure so identity and asset credentials verify by the same rules.
did:aurigraph:...
The compliance layer carries machine-readable proof of which legal frameworks the token satisfies — jurisdiction, applicable framework, and the KYC/AML verification result (level and score). Because it is a discrete layer, a regulator can verify compliance without being exposed to private provenance data, and a buyer can verify the binding without learning the seller's identity. Cryptographic selective disclosure via BBS+ signatures — proving "this token is MiCA-compliant" without revealing the rest of the envelope — is the design target; in the current codebase the BBS+ engine is a development simulation that performs no on-curve cryptography (AV11-3116), so today's privacy guarantee is the layer separation itself, not a zero-knowledge proof.
The provenance layer is the asset's complete origin-and-lifecycle history: custody events, sensor and dMRV readings, appraisals, and every state transition — all linked in a SHA-256 hash chain. This layer is populated and continuously extended through Aurigraph's DMRV (Digital Monitoring, Reporting & Verification) pipeline: every token operation emits a DMRV event to a tamper-evident, consensus-anchored audit chain (see §4). It is the layer that lets an auditor verify a year of transfers in seconds and an insurer price coverage against cryptographically guaranteed history rather than manual paperwork.
The binding layer is what makes a CTF token a claim about this specific asset rather than an abstract unit. It carries the cryptographic link to the artefact — an NFC tag UID, an HUID hallmark on a gold bar, or an IPFS hash of the governing documents — so a verifier can check the credential against the physical thing (an NFC tap, a hallmark lookup) and get a match-or-not answer.
This layer is also where the token's legal terms live. They are not free text. CTF renders an ISO-aligned Ricardian contract into the credential: a document that is simultaneously human-readable (a narrative a court can read) and machine-parseable (structured fields a verifier can evaluate). On Aurigraph this is realized through Active Contracts — library templates whose declarative automation rules mint and transition tokens on state entry, so the legal instrument and the token lifecycle are the same object. The 37 seeded templates map to LADM (land), UBL 2.1 (trade), Dublin Core (documents), IPTC and EXIF across six jurisdiction groups, and Aurigraph's Legal BERT risk assessment flags anomalous deviations before issuance (it degrades to a keyword fallback when the hosted model is unavailable). Status as of 2026-10-20: contracts are signed with ML-DSA-87 (AV11-3761), multi-party signing is being fixed (AV11-3840), and anchoring the signed contract to the ledger, signature-gated activation and the prose + parameters + code triple hash are not yet done (AV11-3863) — so the "Ricardian legal terms" entry in the binding layer is a reference to the stored contract, not yet a ledger-anchored hash.
The key layer holds the signing material and its rotation policy — the ML-DSA-87 public key (multibase / Base58-BTC encoded) and a rotation schedule. Keeping key material as an explicit layer means a token can rotate its keys without re-issuance: when a quantum-capable adversary emerges or NIST standardizes a stronger primitive, the new key signs the next attestation and the verification chain stays intact through the Provenance layer's hash chain.
The five layers are not merely stored side by side. The outer Verifiable Credential proof (§3) signs across all of them, and the Provenance hash chain links each new attestation to the last. The result is an all-or-nothing envelope: alter the compliance layer, strip the binding, or substitute a key, and the single signature check fails. This is precisely the property that separate-but-linked tokens cannot offer.
CTF does not define a new credential format. The five layers are packaged as a single W3C Verifiable Credential — the standard, interoperable envelope for machine-verifiable claims — so any conformant verifier can consume an Aurigraph token.
{ "@context": [ "https://www.w3.org/2018/credentials/v1", "https://w3id.org/security/suites/jws-2020/v1" ], "type": ["VerifiableCredential", "AurigraphCompositeCredential"], "issuer": "did:aurigraph:5Hq3...", "issuanceDate": "2026-04-23T10:30:00Z", "credentialSubject": { "id": "did:aurigraph:asset:RWA-RE-20260221", "identity": { "controller": "did:aurigraph:5Hq3...", "authentication": ["..."] }, "compliance": { "jurisdiction": "US", "framework": "MiCA", "kycLevel": "ENHANCED", "score": 95 }, "provenance": { "merkleRoot": "a1b2c3...", "mlDsa87Signed": true }, "binding": { "huid": "IN-HM-...", "documentHash": "ipfs://Qm..." }, "key": { "type": "ML-DSA-87", "publicKeyMultibase": "z..." } }, "proof": { "type": "DataIntegrityProof", "cryptosuite": "aurigraph-ml-dsa-87-jcs-2026", "verificationMethod": "did:aurigraph:5Hq3...#key-1", "proofPurpose": "assertionMethod" } }
Post-quantum proofs. Every credential — and every Merkle root beneath it — is designed to be signed with ML-DSA-87 (NIST FIPS 204, formerly CRYSTALS-Dilithium), the highest post-quantum security tier (NIST Category 5, comparable to AES-256). Status as of 2026-10-19: ML-DSA-87 Data Integrity proofs on platform-issued credentials are in the V12 source (AV11-3534), and the registry signing key is generated in memory per tree and not yet persisted. Leaf and tree hashing use SHA-256 (FIPS 180-4); keys are encoded in W3C Multibase (Base58-BTC). The design goal is that the entire ownership and provenance chain stays secure against a future quantum adversary — including the "harvest now, decrypt later" threat that already imperils ECDSA/Ed25519-signed assets on public chains.
A verifier's job is therefore reduced to a small, fixed set of checks:
Holder ──── presents VC ───▶ Verifier │ ├─ check ML-DSA-87 signature ─▶ valid? ├─ check binding (NFC / HUID) ─▶ matches artefact? ├─ check compliance scope ─▶ in policy? └─ check provenance hash chain ─▶ unbroken? │ ▼ ✓ VERIFIED (single boolean)
A CTF token is not static; it accrues an append-only event history, and every transition emits a DMRV event that is hash-chained per asset and committed through Aurigraph's HyperRAFT++ consensus engine (which runs in SIMULATE mode on a single node in production today; real multi-node consensus is planned). The credential is re-signed (or a new attestation is anchored) at each step, so the envelope always reflects current state while preserving the full history.
CREATED
COMPOSITE_TOKEN_CREATED
VERIFIED
COMPOSITE_TOKEN_VERIFIED
CONTRACT_BOUND
COMPOSITE_CONTRACT_BOUND
COMPLIANCE_CHECKED
TRANSFERRED
OWNERSHIP_TRANSFER
SETTLED
Merkle anchoring modes. Anchoring is tunable to the asset's risk profile:
Inclusion proofs are O(log n) — roughly 20 hash computations for a registry of up to ~1M entries (Merkle depth 20). The ML-DSA-87 signing primitive has been measured at about 0.3 ms per signature on the production code path (JMH, 2026-10-19, loaded laptop — an upper bound), and DMRV emission is fire-and-forget, so the design intent is that the audit trail imposes negligible cost on the hot path; the end-to-end anchoring latencies have not been benchmarked and are stated here as targets.
O(log n)
CTF is the shared substrate beneath several Aurigraph verticals. In each, the same five-layer credential is specialized with domain-appropriate content in its Binding and Compliance layers.
Across these, the CTF primitive supports fractional ownership, secondary-market transfer, dividend/yield distribution, and redemption — with every event captured in the append-only history of §4.
Cryptographic security.
Post-quantum signatures at NIST Category 5 defend against both live quantum attack and retroactive "harvest now, decrypt later" exposure. Selective disclosure via BBS+ is intended to let a holder prove one layer (e.g. compliance) without revealing the others, satisfying data-minimization requirements, once the real BBS+ implementation replaces the simulation.
Standards & regulatory alignment. The table lists the standards the CTF data model is designed to align with; none of these entries is a certification, audit or regulatory approval.
Because CTF composes these standards rather than replacing them, an Aurigraph token is consumable by any conformant W3C verifier and mappable onto existing regulatory reporting — interoperability over lock-in.
A plain token proves ownership and nothing more — a fatal gap for regulated real-world assets, where identity, compliance, provenance, physical binding, and quantum-durable trust are all prerequisites, not extras. The Composite Token Framework answers this by folding five independently verifiable layers into a single W3C Verifiable Credential, signed with NIST Category 5 post-quantum cryptography and anchored, event by event, to a tamper-evident DMRV chain. The result is a tokenization primitive an auditor, regulator, insurer, or counterparty can verify on first contact with one signature check — and the substrate on which Aurigraph builds gold/HUID, CBDC and trade finance, carbon, and securities tokenization. CTF is patent-pending (PCT-AUR-CTF), and by design it composes existing standards rather than inventing new ones, so the assets it secures interoperate with the world they must live in.
For the platform this framework runs on — HyperRAFT++ consensus, NIST post-quantum cryptography (FIPS 203/204/205), and real-world-asset settlement — see the Aurigraph DLT whitepaper.
Explore: About · Products · Features · Whitepapers · Sustainability · Tokenized carbon credits · Real-world asset tokenization · SDK · Blog · Plain-text site index · Sitemap · Contact: contact@aurigraph.io