Native chain primitives
M-of-N multisig + atomic transactions + on-chain data field + public verifiable ledger. Whatever the chain natively exposes — P2WSH/P2TR, SignerListSet, OP_RETURN.
A technology · applications are built on it
TRUST runs the work a smart contract would do inside attested enclaves, signs with an
M-of-N threshold (threshold Schnorr) that assembles no key in one place, and settles on the
chain’s own native primitives — multisig, atomic tx, on-chain data. Code updates activate only when the
operator quorum signs a new measurement. Phoenix, XPerp and the bridge are applications of this engine;
the next one plugs into the same stack.
// 01 The wedge
Every chain whose settlement layer carries no Turing-complete VM — a meaningful, structurally underserved slice of market cap. Not an EVM/SVM play. On XRPL, TRUST is the intermediate layer: enclave logic today for what the ledger doesn’t do natively — shielded transfers, lending logic, agent-payment policy — and a path to an amendment once it does.
// 02 Composition — three primitives
Mutually reinforcing, not additive. Drop any one and TRUST collapses into something that already exists — a consortium chain, signing-as-a-service, or attested SaaS compute.
M-of-N multisig + atomic transactions + on-chain data field + public verifiable ledger. Whatever the chain natively exposes — P2WSH/P2TR, SignerListSet, OP_RETURN.
Matching, settlement math, threshold signing inside hardware-isolated enclaves, measured and remotely attested. Keys never leave silicon.
Cryptographically enforced. Operators control which code runs without controlling what each operation does — reproducible build → the quorum admits a measurement.
// 03 The stack — technology vs applications
P2WSH / P2TR / threshold Schnorr), EVM (Safe + V7 registry), XRPL (SignerListSet + escrow + crypto-conditions)L0 + L1 are the technology (production / all shipped). L2 is where applications live — each inherits the trust model, attestation pipeline, multisig settlement and operator quorum. Build an app, don’t rebuild the engine.
// 04 Running today — three flagships (L2)
t-of-n SignerList · escrow + crypto-conditionsthreshold Schnorr t-of-n / P2WSH · OP_RETURN commitmentStatus · testnet MVP
Status · testnet, relaunch next
TRUST Bridge
Status · testnet · privacy in development
The technology is the product. Applications are interchangeable.
Each new app reuses L0/L1 and inherits everything; the deltas are measured per target. Confidant is our compliant shielded pool for XRPL — encrypted notes and nullifiers with selective disclosure via viewing keys (hackathon build, testnet). Run its pool inside the enclave and you get shielded XRPL on attested hardware — enclave-enforced privacy alongside the zk design.
Confidant → shielded XRPL
Confidant’s private XRPL pool running inside the enclave — same L0/L1, new app logic.
Options DEX
Same M-of-N settlement on BTC / XRPL, different payoff math.
Lending vaults
Native-coin collateral, enclave-margined; very high framework reuse.
Attested oracles / custody-as-service
API over the ceremony toolchain; cross-chain swaps; BTC↔Monero.
// 05 The differentiator
Most systems decentralize custody but keep a centralized upgrade path — an admin key or a git push that silently changes what every node runs. That path is the one TRUST closes first.
Any change to the enclave code changes its measurement, and the new measurement is
admitted only by a quorum ceremony of the physical nodes — the same t-of-n authority that
governs funds. Builds are reproducible, so independent operators each rebuild to the identical
measurement before they consent. No admin key, no out-of-band push — for every build, test versions included.
// 06 Trust model & trade-offs, declared
What you trust
What you don’t
cancel_after; funds are never quorum-captured.// 07 Not bound to one enclave platform
The reproducible-build invariant carries across hardware platforms: the dev team cannot ship a binary that doesn’t reproduce, and the operator can refuse to run code they didn’t agree to. A per-platform migration analysis lives in-repo. Closest analog overall: a permissioned chain — TRUST adds execution privacy from the operators themselves and chain independence. Validated 2026-06: 3-peer cluster producing valid commits.
One engine. Any application. No single party that can betray it.
Cross-chain bridge · Security by construction
The TRUST Bridge — an application of the TRUST engine — is a cross-chain bridge built so that no single key, server, or operator can move funds — because that single point is exactly what attackers have taken, again and again.
Why this matters
Attackers obtained a majority of the validator keys. The bridge required a threshold of signers, but too few independent parties actually held them — so one intruder reached the majority alone.
class: a threshold on paper, not in independent hands“Multi-party” in name, one party in fact: every signing key of the bridge sat with its CEO. When he was arrested, the keys went with him — funds moved without authorization, nobody could stop it, and the bridge shut down.
class: a custodian, not a protocolA consensus bug minted wrapped bitcoin that nothing backed. The federation's threshold signed the ~4,000 BTC peg-out exactly as designed — no key was stolen; the signers released funds against a claim that was never proven locked.
class: the threshold signed what consensus said, not what was provenDifferent bugs, one root: trust concentrated in a few keys, in one custodian, in a single check, or in a claim nobody proved. TRUST is designed to remove that root — in three ways.
The core — threshold everywhere
Authority over funds is split into n shares held by separate, independent nodes. Any t can jointly authorize a transfer; fewer than t cannot — and learn nothing about the key.
An attacker breaks into one node. It holds a single share — which signs nothing and reveals nothing.
To move any funds they need t independent parties at the same moment. No single breach, bribe, or bug reaches that line.
That line is what Ronin crossed: the threshold existed, but too few independent hands held the keys. TRUST's rule is that the parties must be genuinely independent.
Two more defenses
The software guarding the key shares runs inside hardware-secured enclaves. Each node's exact code is cryptographically measured and checked before it is trusted; keys are never exposed in plaintext — not even to the operator.
A transfer is accepted only when proven by the origin chain's own consensus (a light-client / SPV proof), and the amount released is tied to what was proven locked at the source. This is the check Wormhole was missing — nothing can be minted without real backing.
Built on Zcash's shielded (Orchard) technology so value can move through the bridge — between any two chains the engine connects — without exposing amounts and counterparties. The cryptography is well established; the capability is in development, with BTC ↔ Zcash as the first shielded direction.
A note on direction
A shielded bridge is any-to-any. BTC ↔ Zcash is one direction — the first in development. Privacy is a property of the bridge, not of one pair of chains: the same shielded path can be constructed between any two chains the engine connects. BTC ↔ Zcash comes first because it shows a move no conventional bridge can make — across chains with no native VM to verify each other, with amounts and parties kept private (shielded path in development). The technology earns its place where a chain can't verify the other itself — not an EVM-to-EVM hop, where on-chain contracts already do that.
The contrast
Every common bridge design keeps a single point that must be trusted. TRUST replaces each with something you verify.
Honest status
Running and measured
Designed, mathematically grounded, not yet live
TRUST currently runs on testnet; the parameters above describe the current design and testnet deployment. Threshold cryptography protects the integrity of funds — who can move them; the shielded-transfer work adds privacy, and it is still in development.
Replace “trust the operator” with “verify the mathematics.”
TRUST — threshold-secured · attested · proof-checked