@maiat/viem-guard v0.2.0

Agentic trust layer for viem — trust scoring, anti-poisoning, TrustGateHook integration, and threat reporting.

Install

npm install @maiat/viem-guard viem

Quick Start

import { createWalletClient, http, parseEther } from 'viem'
import { base } from 'viem/chains'
import { withMaiatTrust } from '@maiat/viem-guard'

const client = withMaiatTrust(walletClient, {
  minScore: 60,
  antiPoison: true,       // detect address poisoning
  reportThreats: true,    // auto-report to Maiat network
})

await client.sendTransaction({
  to: '0xSomeContract',
  value: parseEther('1'),
})

Agent Wallet (Privy / EIP-1193)

One-line integration for AI agent wallets:

import { createMaiatAgentWallet } from '@maiat/viem-guard'

const wallet = createMaiatAgentWallet(privyProvider, {
  minScore: 70,
  antiPoison: true,
  apiKey: 'mk_...',
})

// All transactions: trust-gated + anti-poisoned + threat-reported
await wallet.sendTransaction({ to, value })

TrustGateHook Integration (Uniswap V4)

Fetch EIP-712 signed scores for TrustGateHook-protected pools:

import { fetchSignedScore, encodeSwapHookData } from '@maiat/viem-guard'

// Fetch signed trust scores for both tokens
const score0 = await fetchSignedScore('0xToken0Address')
const score1 = await fetchSignedScore('0xToken1Address')

if (score0 && score1) {
  const hookData = encodeSwapHookData(myAddress, score0, score1)
  // Pass hookData to your Uniswap V4 swap transaction
}

Anti-Poisoning Engine

Detects address poisoning attacks before they happen:

• Vanity Match Detection: Scans wallet history for addresses with matching first4+last4 hex chars
• Liveness Check: Flags accounts created <24h ago with dust-only transactions

const client = withMaiatTrust(walletClient, {
  antiPoison: {
    vanityMatch: true,
    livenessCheck: true,
    etherscanApiKey: 'YOUR_KEY',
    explorerUrl: 'https://api.basescan.org/api',
  },
})

Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | minScore | number | 60 | Block if trust score below threshold | | apiKey | string | - | mk_... key for paid tier | | mode | 'block' \| 'warn' \| 'silent' | 'block' | How to handle low-trust | | onWarn | (result) => void | - | Called in warn mode | | recordOutcomes | boolean | false | Record tx outcomes to Maiat | | antiPoison | boolean \| AntiPoisonConfig | false | Enable anti-poisoning | | reportThreats | boolean | true | Auto-report threats to network | | routerAddresses | string[] | - | Uniswap V4 router addresses | | feeTarget | string | - | Address for fee discount |

Error Handling

import { withMaiatTrust, MaiatTrustError, MaiatPoisonError } from '@maiat/viem-guard'

try {
  await client.sendTransaction({ to: '0x...', value: parseEther('1') })
} catch (e) {
  if (e instanceof MaiatPoisonError) {
    console.log(e.threatType)      // 'vanity_match' | 'dust_liveness'
    console.log(e.matchedAddress)  // the real address being impersonated
  }
  if (e instanceof MaiatTrustError) {
    console.log(e.score)           // 0-100
    console.log(e.verdict)         // 'block'
  }
}

Collective Immunity

When Guard blocks an attack, it automatically reports the malicious address to the Maiat network (privacy-safe — no sender context). Every Maiat-protected agent gets instant protection.

Agent A blocks attack → Report to Maiat → Global TrustScore update → All agents immunized

Exports

// Core
export { withMaiatTrust } from '@maiat/viem-guard'
export { createMaiatAgentWallet } from '@maiat/viem-guard'

// Hook Data
export { fetchSignedScore, encodeSwapHookData } from '@maiat/viem-guard'

// Anti-Poisoning
export { detectVanityMatch } from '@maiat/viem-guard'

// Threat Reporting
export { reportThreat } from '@maiat/viem-guard'

// Trust Check
export { checkTrust } from '@maiat/viem-guard'

// Types & Errors
export { MaiatTrustError, MaiatPoisonError } from '@maiat/viem-guard'
export type { MaiatCheckResult, SignedScore, AntiPoisonConfig, ThreatReport } from '@maiat/viem-guard'

Powered by

Maiat Protocol — trust infrastructure for agentic commerce