HashKey Chain Testnet

The AI-Powered Risk & Compliance Layer for HashKey Chain

4-layer AI scoring pipeline. On-chain enforcement. Any DeFi protocol integrates in 2 lines of Solidity.

API Response

// GET /v1/risk/0x2222...2222
{
  "wallet": "0x2222...2222",
  "score": 8,
  "level": "HIGH",
  "flags": ["mixer_exposure"],
  "isCompliant": false,
  "hopDistance": 1,
  "reasoning": "Direct mixer contact"
}

Features

Everything you need to enforce compliance

AI-powered risk scoring, on-chain enforcement, and composable integration — all built natively for HashKey Chain.

Wallet Risk Scoring

4-layer AI pipeline scores every wallet 1-10. Attribution, sanctions, graph analysis, and AI behavioral enhancement.

Sanctions Screening

Real-time OFAC SDN screening. Sanctioned addresses are instantly flagged as CRITICAL (score 10) and blocked.

Payment Risk Assessment

End-to-end A-to-B payment risk checks. Both sender and receiver scored. Returns block, flag, or allow recommendation.

On-Chain Enforcement

AegisGuard modifier blocks non-compliant wallets automatically. No off-chain middleware. Pure on-chain enforcement.

Graph Proximity Analysis

Hop distance methodology traces connections to known malicious actors. 1 hop = HIGH risk, 4+ hops = LOW risk.

Composable by Design

2 lines of Solidity to integrate. Inherit AegisGuard, add onlyCompliant modifier. Works with any DeFi protocol.

How It Works

From wallet to enforcement in 3 steps

AI Scores Every Wallet

4-layer pipeline: Attribution override, OFAC sanctions screening, graph proximity analysis, and Groq AI behavioral enhancement.

Score Written On-Chain

Risk score (1-10), flags, reasoning, and hop distance are written to AegisRiskScore.sol on HashKey Chain. Cached in Neon DB.

Enforcement Is Automatic

DeFi protocols inherit AegisGuard. The onlyCompliant modifier blocks non-compliant wallets. No custom logic needed.

Scoring Pipeline

4 layers of AI-powered risk analysis

Each wallet passes through 4 sequential layers. Early layers can short-circuit — sanctioned wallets never reach the AI layer.

Attribution

Known clean addresses are instantly whitelisted with score 1. Prevents false positives on verified infrastructure.

Result

Score 1, STOP

Sanctions

Real-time OFAC SDN screening. Any match = immediate score 10 CRITICAL. No override possible.

Result

Score 10, STOP

Graph Analysis

Traces hop distance to known malicious actors. 1 hop = HIGH risk, 4+ hops = LOW risk. Last 2000 blocks.

Result

Base score

AI Enhancement

LLM detects behavioral patterns: rapid movement, dormant activation, clustering. Adjusts score ±2 points.

Result

Adjust ±2

Final Score

final = (graph × 0.6) + (AI × 0.4)

Clamped 1-10, written on-chain to AegisRiskScore.sol

For Developers

2 lines of Solidity. That's it.

Inherit AegisGuard and add the modifier. Your protocol now has AI-powered compliance enforcement.

MyDEX.sol

import "./AegisGuard.sol";

contract MyDEX is AegisGuard {
  constructor() AegisGuard(AEGIS_ADDRESS) {}

  function deposit()
    external payable
    onlyCompliant  // <-- this is it
  {
    // your logic here
  }
}

REST API

// Read risk profile
GET /v1/risk/:wallet

// Run full AI scoring pipeline
POST /v1/risk/analyze
// body: { "wallet": "0x..." }

// Payment risk A → B
POST /v1/risk/payment
// body: { "from": "0x...", "to": "0x..." }

// Scoring audit trail
GET /v1/risk/:wallet/history

Score System

Risk levels & enforcement

Two enforcement levels: strict (onlyCompliant) blocks HIGH+, soft (onlyNotCritical) blocks CRITICAL only.

Score	Level	Compliant	deposit()	swap()
10	CRITICAL	No	Blocked	Blocked
7 - 9	HIGH	No	Blocked	Allowed
4 - 6	MEDIUM	Yes	Allowed	Allowed
1 - 3	LOW	Yes	Allowed	Allowed