Web3 has mastered verifiability, programmability, and permissionless access. What it still lacks is interpretability—a way to understand the reliability of participants who interact with decentralized systems. Blockchains record transactions with perfect fidelity, yet they offer no semantics around trust, reputation, or behavioral quality. As a result, crypto has resorted to blunt economic tools: overcollateralized lending, blanket airdrops, restrictive allowlists, and endless attempts at Sybil resistance.

zScore introduces a new primitive: a universal, AI-based, verifiable reputation score for wallets, secured through an EigenCloud AVS and propagated across ecosystems via LayerZero. It converts fragmented activity across chains into a coherent, cryptographically verifiable reputation state that can be used by protocols, agents, and applications.

This article presents a balanced, research-style overview of how zScore works, how the AVS architecture ensures neutrality, and how reputation reshapes both protocol design and user experience.

Crypto’s data abundance masks a deeper problem: chains provide raw signals but no interpretation. A wallet that behaves responsibly during market stress looks identical—at the data level—to a bot engaging in repetitive interactions.

The absence of reputation has created systemic inefficiencies:

• Airdrops leak 50–80% of value to opportunistic wallets

• Lending markets require 150–200% collateral because all borrowers appear equally risky

• NFT mints are botted, forcing projects to centralize allowlists

• Governance is Sybil-prone, despite the presence of long-term contributors

• AI agents lack safety rails, since nothing distinguishes a reliable agent from a malicious one

zScore addresses these issues by extracting behavioral meaning from onchain actions, producing a reputation score based entirely on public, non-identity data.

zScore operates on a simple premise: wallet behavior is economically meaningful.

Its neural network analyzes:

• Transaction patterns and historical behavior

• Lending/repayment reliability

• LP consistency and liquidity provisioning dynamics

• Staking and restaking stability

• Airdrop and governance participation

• Multi-chain sophistication

• Responses to volatility

• Indicators of Sybil, bot, or farm-like activity

The output is a 0–1000 reputation score. But critically, this score doesn’t stand alone. Instead, the system produces a Merkle tree, where:

• Each wallet’s score forms a leaf

• The tree compresses the global reputation set

• The Merkle root becomes the canonical “state commitment”

This structure enables lightweight verification: anyone can present a score + Merkle proof, and a smart contract can confirm that it is part of the global attested state.

A reputation system cannot rely on a single party. If Zeru alone published its own Merkle root, the ecosystem would be asked to trust the entity generating the data—a contradiction to Web3 principles.

This is why zScore is built on the top of EigenCloud infrastracture.

• Decentralized verification: Operators independently validate the Merkle root

• Economic incentives: Operators stake and restake capital, placing economic weight behind correctness

• Separation of roles: Zeru computes the tree; operators attest to its validity

• Credible neutrality: No single actor can manipulate reputation

1. Zeru publishes the proposed Merkle root

2. Operators run the zScore Attester, recomputing/validating the state

3. Operators reach consensus and sign the commitment

4. The Othentic Aggregator pushes the verified root onchain

This process converts reputation from a subjective assessment into an operator-backed attested fact—a new form of truth in decentralized systems.

Reputation loses power if fragmented. In a multi-chain world, a wallet’s behavior on one chain must inform its treatment on all others.

LayerZero enables this by transporting the operator-verified Merkle root to every supported EVM chain.

• All chains share the same reputation state

• Protocols can rely on consistent truth, regardless of where they run

• AI agents and smart wallets operate with the same constraints everywhere

• Governance, lending, and DeFi systems gain a unified behavioral oracle

The result is a cross-chain reputation layer that sits above individual L1s and L2s, shaping incentives across the entire ecosystem.

The implications of zScore extend far beyond scoring.

Lending markets can finally price risk dynamically rather than assuming all borrowers are equal. High-reputation wallets receive better LTVs and lower rates, while risky wallets face safeguards. This unlocks undercollateralized credit—long considered impossible in Web3.

Projects can target real contributors instead of bot clusters. Reputation-weighted distributions reduce spillage, increase retention, and fundamentally shift airdrops from “marketing expense” to “capital allocation.”

DEXs and liquidity protocols can offer structural advantages to reliable LPs and traders:

• reduced fees,

• preferential routing,

• safer pools,

• more efficient incentives.

DAOs can prioritize long-term wallets over extractive ones.
Reputation becomes a mechanism for ensuring deliberation, not just voting power.

Delegator reputation introduces a new dimension of safety and predictability to restaking ecosystems.

As agents begin executing transactions autonomously, reputation becomes a safety primitive:

• bounding agent permissions,

• setting spending limits,

• ensuring predictable execution,

• preventing systemic abuse

In short: reputation becomes a first-class economic variable.

While zScore functions as an infrastructural primitive, zPass transforms it into a user product: a universal reputation passport minted once and recognized everywhere.

• A persistent representation of one’s onchain reliability

• Better access to fees, rates, mints, allocations, and rewards

• A portable behavioral history

• A visual identity (Houses, Tiers, Status)

• A unified layer of trust across protocols

Users earn benefits by being reliable, not by revealing identity.
zPass turns behavior into a long-term asset, aligning users and protocols in a healthier way.

The emergence of a decentralized reputation system marks a structural evolution in Web3:

• Phase 1: Transparent transactions

• Phase 2: Programmable finance (DeFi)

• Phase 3: Shared security (restaking → EigenCloud)

• Phase 4: Interpretability — systems that understand participants

Part of what we’re doing here at Zeru is making all of this data accessible and interpretable by the average person. zScore, as a value between 0 and 1000, along with attributes like wealth, consistency and protocol diversity makes it much easier for users to view and understand what the impact of their onchain behavior leads to.

Another part of making this data accessible is opening up the zScore Agent on Sentient, whose mission is to ensure that Artificial General Intelligence is open-source and not controlled by a single entity. You will soon be able to chat with the zScore model and ask it for the zScore of any wallet, for insights on that wallet’s behavior, analyze protocols, and more. Offering this in a familiar chat interface will open up possibilities never seen before!

Reputation closes the loop between behavior and incentives. It strengthens markets, stabilizes ecosystems, enhances user experience, and provides a safety layer for increasingly autonomous web economies.

zScore and zPass represent the first practical implementation of this vision:
an interpretable, verifiable, universally accessible reputation system built directly into the multi-chain fabric of Web3.

As crypto moves toward an AI-native future, the need for such a layer becomes not merely beneficial, but inevitable.