FETCH.AI · L1 · STAGE 0 UNAWARE · QRI 22 v3.1.0 methodology

Fetch.ai is a Cosmos-SDK / Tendermint L1 (chain-id fetchhub-4) using Ed25519 at validator consensus and secp256k1 ECDSA at the account layer, with no published PQC roadmap on the Fetch L1 itself. The ASI Alliance's planned ASI:Chain (separate future L1, testnet 2026, mainnet late 2026 / early 2027) has been described as having a modular cryptography layer that could plug in lattice or hash-based primitives, but that is a different chain and not yet in production. uAgent / Agentverse / DeltaV agent-identity signing relies on the same classical primitives and inherits their quantum vulnerability.

inLinkedIn ↷Audit access ⇆Compare Verified 2026-05-02

Summary

Fetch.ai scores QRI 22, Band 3 Planning, Migration Stage 0. The L1 (fetchhub-4) is a Cosmos-SDK fork with Tendermint consensus: Ed25519 validator signing, secp256k1 ECDSA account signatures, SHA-256 hashing, X25519/Ed25519 for the Tendermint p2p secret-connection handshake, classical TLS for RPC/REST. Fetch's differentiators (Distributed Random Beacon integration with Tendermint via cosmos-consensus, uAgents framework, Agentverse, DeltaV, AI-Engine, CosmPy) all build on the same classical primitive base. The active validator set was raised to 60 (governance proposal), smaller than Cosmos Hub's ~180. No PQC ADR, no PQC governance proposal, no PQC milestone on Fetch.ai L1; recent governance attention sits on the ASI token-merge ($FET + $AGIX + $OCEAN → $ASI) and ecosystem expansion. ASI Alliance discussion of post-quantum primitives (lattice + hash-based) targets the future ASI:Chain, not the Fetch L1, and remains a roadmap statement rather than shipped code. Mainnet-Traffic cap binds at 5a=0%. Architecture-Execution Gap cap also binds (Dim 4 ≫ Dim 5). No agent-platform-specific PQC attestation flow declared.

What the gates say

Gate 1a, Hybrid signature: FAIL , no documented hybrid signature composition AND or OR on Fetch.ai L1; no FIP, no roadmap, no spec for hybrid Ed25519+PQ or secp256k1+PQ
Gate 1a, Hybrid KEM: FAIL , Tendermint secret-connection X25519, RPC TLS classical X25519/RSA; no hybrid PQ KEM deployed; uAgent message channels classical
Gate 1b, Commit-to-hash: COND , only relevant if 1a-Sig passes via OR-composition
Gate 2, Evidence reconstruction: PASS , every sub-score reconstructible from public artifacts within 48 hours
Gate 3, Primitive naming: PASS , primitives named at every sub-score

Burn-vs-rescue policy on file

Declared option f, Undeclared. No published Fetch.ai policy on what happens to FET / ASI at quantum-vulnerable accounts post-CRQC. No freeze/burn proposal, no STARK rescue scheme, no rate-limit canary, no client-layer hybrid migration framework. Agent-wallet posture also undeclared.

Seven dimensions

Each dimension scores 0–100 internally; the weighted roll-up produces the QRI.

1 Cryptographic Exposure weight 15% 26 / 100

1a · primitive inventory 13 / 20

Inventory clear and reconstructible from fetchai GitHub (fetchd, cosmos-sdk fork, cosmos-consensus). uAgents identity reuses Cosmos account-key primitives.

Primitives: Ed25519 (Tendermint validator consensus signing, default in fetchai/fetchd and fetchai/cosmos-consensus) · secp256k1 ECDSA (Cosmos SDK account signatures via fetchai/cosmos-sdk fork, R||S lower-S form, SHA-256 message digest) · SHA-256 (block hashing, IBC commitment hashing) · X25519/Ed25519 (Tendermint p2p secret-connection handshake) · Standard TLS for RPC/REST endpoints · DRB threshold/pairing primitives (Distributed Random Beacon integrated via fetchai/cosmos-consensus)

1b · shor grover pq tag 5 / 20

No PQ-safe primitive in active use on Fetch.ai L1.

Tags:

Ed25519 → Shor-break-via-DL-without-pairings
secp256k1 ECDSA → Shor-break-via-DL-without-pairings
SHA-256 → Grover-weaken (256→128-bit)
DRB pairing/threshold primitives → Shor-break-via-pairings (where pairing-based) or Shor-break-via-DL (threshold over EC)

1c · family diversity 0 / 20

0 PQ families. Two classical families on the signing path (Edwards-curve EdDSA, Weierstrass-curve ECDSA), neither PQ-safe.

1d · nist security category 2 / 20

Ed25519 ≈ 128-bit classical / 0-bit post-Shor; secp256k1 ECDSA ≈ 128-bit classical / 0-bit post-Shor; SHA-256 ≈ 128-bit post-Grover. No NIST PQC primitive in scope.

1e · implementation quality 6 / 20

Tendermint consensus has TLA+ specification work (upstream); Fetch's cosmos-consensus DRB integration adds custom logic without machine-checked PQ-relevant proofs. Standard Go crypto/ed25519 (constant-time); secp256k1 via btcec. Tier 1 (mature classical EC + SHA-2). No PQ implementation, no formal verification of the DRB at the level of Formosa-Crypto / Hülsing XMSS proofs.

2 Quantum Recovery Exposure weight 10% 19 / 100

Forge subtotal: 14/75 Decrypt subtotal: 5/25

2a · active key exposure 4 / 25

fetch1… accounts derive from secp256k1 pubkey hash (Cosmos SDK pattern). Pubkey is published on-chain on first outbound tx; any account that has ever signed has its pubkey publicly recorded, Shor-forgeable post-CRQC. uAgent wallets follow the same pattern and broadcast pubkeys when paying in FET / ASI.

2b · cold key exposure 5 / 25

Accounts that have never signed retain pubkey-hash protection; accounts that signed once retain exposed pubkeys indefinitely. fetchhub mainnet has been live since 2019 (fetchhub-1 → fetchhub-4 lineage), so a non-trivial fraction of cold supply has revealed pubkeys at some point. No public quantification.

2c · sig long term validity 5 / 25

Every historical Ed25519 validator vote and secp256k1 account signature is forgeable after CRQC. Agent-economic provenance (uAgent reputations, AI-Engine / DeltaV service-agreement signatures) is signed under the same classical keys and is therefore retroactively forgeable. IBC light-client trust on Fetch's IBC channels extends this to counterparties.

2d · encryption confidentiality hndl 5 / 25

Tendermint p2p secret connection uses X25519 ECDH for transport encryption between validators (Shor-vulnerable). RPC/REST endpoints use standard TLS (classical X25519 / RSA / ECDH). uAgent message channels rely on the same classical KEMs. No hybrid PQ KEM deployed on Fetch.ai L1 or in the agent stack.

3 Metadata, Anonymity & Confidentiality weight 13% 19 / 100

3a · tx graph visibility 5 / 20

Fully transparent ledger; fetch1… addresses pseudonymous; agent-to-agent payment flows on-chain are linkable. Agentverse-hosted agents add an off-chain metadata surface (mailbox, message queue) that is not on-chain but is observable to the platform operator.

3b · rpc mempool concentration 4 / 20

Public RPC endpoints concentrated among Fetch.ai-operated infrastructure plus a handful of community providers (AviaOne, Lavender.Five, Polkachu). Smaller validator set (~60 cap) plus Fetch-foundation hosted endpoints means greater concentration than larger Cosmos chains. No validator-metadata-retention policy declared.

3c · cross chain bridge correlation 3 / 20

Fetch is IBC-connected; flows between fetchhub and other Cosmos zones are directly linkable. ASI Alliance token-merger cross-chain mechanics (FET + AGIX on Ethereum + OCEAN on Ethereum → ASI) introduce additional EVM-side correlation.

3d · retroactive de anonymization 7 / 20

Fetch.ai L1 does not publish encrypted payload data, ZK-shielded transactions, or DL-based ring signatures at the protocol level. Some agent-to-agent message payloads may carry confidential commercial content protected by classical PK schemes, these would be HNDL-decryptable post-CRQC, but they are an off-chain layer and not part of the consensus surface.

3e · mixnet shuffle 0 / 20

None at protocol level. Agent-to-agent traffic in Agentverse / Mailroom is not a cryptographic mixnet.

4 Migration Architecture weight 10% 49 / 100

4a · crypto agility 5 / 15

Inherits Cosmos SDK / CometBFT modularity (--key-type flag, modular crypto/keys package). Fetch's cosmos-consensus repo demonstrates capability to fork and modify the Tendermint crypto path (DRB integration). However, no production instance of a validator-consensus-key-type swap on fetchhub mainnet within 5 years; agility is architectural, not demonstrated.

4b · aa key rotation 5 / 20

Cosmos SDK ADR-016 consensus-key-rotation inherited at the SDK level; whether Fetch's older SDK fork has merged the v0.52 Olympus x/accounts module is unverified. x/authz and x/feegrant available via SDK fork. No native account abstraction comparable to ERC-4337 / EIP-7702. uAgents provides agent-level identity but is an application-layer construct, not a chain-level AA primitive. No documented client-layer PQ migration path.

4c · hard fork track record 9 / 15

fetchhub-1 → fetchhub-2 → fetchhub-3 → fetchhub-4 chain-id transitions executed via coordinated upgrades. Validator-set governance proposal (50 → 60 active validators) ratified. ASI token-merge ($FET + $AGIX + $OCEAN → $ASI) executed via on-chain governance. Cadence acceptable but slower / less frequent than Cosmos Hub.

4d · hybrid deployment readiness 4 / 15

Architecturally, the Cosmos SDK + Tendermint base permits hybrid validator-key constructions; Fetch's existing cosmos-consensus fork shows willingness to modify the consensus crypto path. No spec proposal or FIP for a hybrid Ed25519+PQ scheme on Fetch.ai L1. ASI Alliance commentary about modular PQ primitives is scoped to the future ASI:Chain, not fetchhub.

4e · stateful hash state management 15 / 15

N/A by default, no stateful hash scheme in scope; stateless schemes score full per v3.1 rubric.

4f · bft aggregation path 0 / 20

N/A. Tendermint / CometBFT (Fetch fork) uses Ed25519 non-aggregating signatures at consensus. BLS aggregation is not in the Fetch.ai consensus path; the cosmos-consensus DRB uses pairing-based threshold cryptography in a separate role (randomness beacon), not for vote aggregation. Per v3.1 rubric, 4f is N/A for non-aggregating-signature consensus and weight redistributes.

5 Deployment Execution weight 22% 11 / 100

5a · mainnet pqc traffic pct 0 / 25

0% of validator votes or account signatures on fetchhub-4 mainnet under a PQC primitive.

5b · pqc code in consensus client 0 / 15

No PQC scheme merged into fetchai/fetchd, fetchai/cosmos-sdk fork, or fetchai/cosmos-consensus main. No PQ research fork comparable to DoraFactory's tendermint-pqc has been published from Fetch.ai engineering.

5c · validator pqc key adoption 0 / 15

All ~60 active fetchhub-4 validators use Ed25519 consensus keys per Tendermint default. No validator has registered a PQC consensus key.

5d · published dated milestones 0 / 10

VOIDED to 0 per v3.1 rule (5a = 0). No dated, enforcement-mechanism-backed PQC milestones published for Fetch.ai L1. ASI:Chain testnet/mainnet dates are for a separate future chain, not the live Fetch L1, and are not protocol-enforced milestones for fetchhub.

5e · pqc washing delta 9 / 15

Announced PQC trailing-12-mo: ASI Alliance public commentary referencing modular quantum-safe primitives for the future ASI:Chain (not Fetch.ai L1). Shipped PQC: 0. Low-volume narrative; not aggressive washing but creates scope confusion between Fetch.ai L1 (current chain, no PQ work) and ASI:Chain (future chain, PQ statements). 3-point deduction reflects the conflation risk.

5f · signature footprint multiplier 0 / 20

No PQ deployment, no published bytes-per-block analysis under any PQ scheme for Fetch.ai. Undisclosed.

6 Supply Chain Vendor Readiness weight 22% 7 / 100

6a · wallet 2 / 25

Top wallets supporting FET / ASI: ASI Alliance Wallet (Fetch native), Keplr, Leap, Ledger HW, MetaMask (for ERC-20 ASI). None publish a PQC roadmap. Ledger has internal PQC research at Ledger Donjon but no shipped PQ-signing for FET / ASI accounts.

6b · bridge 2 / 25

Top bridges: IBC (light-client, Ed25519 verification), Axelar, Gravity Bridge or equivalent for FET ↔ Ethereum (relevant given ASI is also an ERC-20 on Ethereum). None publish a PQC roadmap. The ASI token-merge cross-chain swap mechanics rely on classical signatures.

6c · custodian 1 / 25

Institutional custody for FET / ASI: Coinbase Custody, Kiln (staking), BitGo, Anchorage exposure varies. None publish a Fetch-specific PQC roadmap. None have MPC-PQ in production for FET / ASI signing. ATOM ecosystem MPC-PQ work would need to be ported.

6d · rpc hsm tee infra 2 / 25

RPC providers: Fetch.ai-hosted endpoints (browse-fetchhub.fetch.ai, explore-fetchhub.fetch.ai), AviaOne, Lavender.Five, Polkachu, community validators. None publish PQ-enabled RPC TLS. HSMs used by validators: standard YubiHSM2 / Ledger / Thales / cloud KMS, no PQ signing for Ed25519/secp256k1 in production. TEE attestation chains not declared in the Fetch validator stack at protocol level. Agentverse cloud infrastructure attestation flow not mapped to PQC.

7 Governance & Coordination weight 8% 38 / 100

7a · validator stake distribution 8 / 20

~60 active validators (raised from 50 via on-chain proposal). Smaller set than Cosmos Hub (~180). Concentration metrics published less frequently; Nakamoto coefficient estimated in low single digits historically. Client diversity weak, single Tendermint/CometBFT-derived consensus client (fetchd / cosmos-consensus).

7b · upgrade cadence under pressure 10 / 20

fetchhub-1 → fetchhub-2 → fetchhub-3 → fetchhub-4 chain-id transitions completed; ASI token-merge governance executed. No clear public record of an emergency security upgrade under attacker pressure comparable to Cosmos Hub v19.2.

7c · named coordination lead 10 / 20

Fetch.ai Foundation / Fetch.ai Ltd. (Cambridge), plus ASI Alliance (Fetch.ai + SingularityNET + Ocean Protocol, post-merger). Clear named ownership at the chain level. No named PQC migration lead for Fetch.ai L1; ASI Alliance public PQ commentary scoped to ASI:Chain.

7d · adversarial coordination precedent 10 / 20

Successful execution of contested governance items (validator-set increase, ASI token-merge proposal). No precedent of a coordinated cryptographic-primitive change under attacker pressure.

7e · canary tripwire mechanism 0 / 20

No canary, honeypot, rate-limited spending rule, or cryptographic tripwire on Fetch.ai L1 or in the agent stack.

X + Y vs Z, when does the math turn against you?

v3.1 demotes the X+Y vs Z timing test to a secondary signal, the headline output is Migration Stage. The timing test still answers the question: can this chain finish migrating before the threat lands?

X, signature shelf life

5–15 years, historical Ed25519 validator votes and secp256k1 account signatures sit on-chain indefinitely; agent-identity signatures issued by uAgents retain validity over the lifetime of agent reputation records

Y, migration time

5–10 years to Stage 5, given Stage 0, hybrid FIP + cosmos-consensus / fetchd merge + 60-validator coordinated rotation + uAgent / Agentverse / DeltaV signing-stack alignment is multi-year minimum; ASI Alliance governance complexity (3-token-merge entity) adds coordination drag

Z10 (10% CRQC year)

2030

Z25 (25% CRQC year)

2035

Verdict

X+Y > 2035, Outside risk window vs Z25; X+Y > 2030, Crisis Zone vs Z10

Z-compliance

Outside compliance window, FET / ASI faces the 2035 disallowance for any classical PK still in active validator-key or account-signing scope

Source-disagreement disclosure

v3.1 requires every chain card to publish material divergences among authoritative sources, plus the delta-QRI under alternative weighting.

Fetch.ai L1 vs ASI:Chain scope

Public ASI Alliance commentary about modular post-quantum primitives (lattice + hash-based) refers to ASI:Chain, a separate future L1 with MeTTa smart-contract language, planned testnet 2026 / mainnet late 2026–early 2027. Card scores Fetch.ai L1 (fetchhub-4) as currently operational, not ASI:Chain claims about a future chain.

Token-merge governance complexity

Post-merger, FET trades as ASI; chain-level governance is shared with SingularityNET and Ocean stakeholders within the ASI Alliance. A coordinated PQC migration on Fetch.ai L1 must therefore navigate alliance-level coordination on top of chain-level validator coordination.

Distributed Random Beacon

Fetch.ai integrates a Distributed Random Beacon (DRB) with Tendermint via the fetchai/cosmos-consensus repo. The DRB itself uses classical pairing-based / threshold cryptography and would inherit quantum vulnerability under any post-CRQC analysis.