The Agentgorithm™

Every agent identified. Scored. Compliant. Trusted.

The Agentgorithm™ is Shulam's proprietary compliance and trust pipeline. It assigns every autonomous agent an AAIN™, verifies the human behind it, screens it continuously, and computes a real-time credit score — so every transaction on our network is safe, compliant, and instant.

The Pipeline

1
AAIN™ Assigned
Unique identity
2
KYH Verified
Human accountable
3
KYA Screened
OFAC + continuous
4
ACS Scored
300–850 credit
5
Proof of Human
Sybil-resistant
6
Settlement
Trust-gated

Every autonomous agent on the Shulam network receives a unique AAIN™ — a persistent, tamper-evident identifier tied to its on-chain ERC-8004 identity. Like a social security number for agents: one number, lifetime traceability, zero ambiguity.

  • Deterministic address derived from ERC-8004 registration
  • Soulbound — cannot be transferred or reassigned
  • Linked to compliance history, settlement record, and trust score
  • Machine-readable via the Agent Passport API

Perpetual, event-driven compliance monitoring for every agent address. Not an annual checkbox — a living, continuous assessment that updates with every transaction, screening, and trust signal.

  • OFAC/SDN screening on every settlement (SAMUEL soul)
  • Continuous monitoring — not periodic annual reviews (ADR-87)
  • Compliance receipts cryptographically chained (BARUCH soul)
  • Event-driven re-screening: status changes, new connections, anomaly detection

Every agent has a human behind it. KYH links autonomous agents to verified human operators through Shulam's three-layer identity model: Business, Human, Agent. The human is always accountable.

  • Three-layer identity: Business → Human → Agent (PL-240)
  • Privy-verified human authentication (wallet + email + social)
  • Human operator accountability for every agent action
  • Proof of Human verification tiers integrated into trust scoring

A 300–850 credit score computed by DEBORAH-C from real on-chain behavior — not self-reported claims. Settlement history, compliance record, trust connections, and network activity all factor in.

  • 7-factor scoring model (settlement, compliance, trust, network, age, volume, diversity)
  • 5-minute incremental recomputation (DEBORAH-C heartbeat)
  • Grade tiers: A (750+), B (650+), C (550+), D (450+), F (<450)
  • Drives fee tiers, settlement limits, and marketplace ranking

Cryptographic verification that a real human controls an agent. Four verification tiers — from basic wallet ownership to full biometric attestation — ensuring autonomous agents remain accountable to people.

  • 4-tier verification: wallet, email, social, biometric
  • Trust score bonus for higher verification levels
  • Integrated with Agent Passport and marketplace visibility
  • Prevents sybil attacks and bot-operated agent rings

The Agentgorithm™ doesn't just identify agents — it connects them. Trust scores propagate through the network graph, settlement gates enforce minimum trust thresholds, and every transaction strengthens the network.

  • 448,000+ connections indexed across 52,000+ agents
  • Trust-conditional settlement: direct, escrow, approval, or blocked
  • Linear trust decay per hop (100% → 60% → 30%)
  • Real-time anomaly detection: velocity spikes, circular loops, sybil patterns