AATS — Agent Attestation Trust Score | Real-time behavioral scoring for autonomous AI agents | 5 scoring dimensions: Identity Integrity (30%), Behavioral Consistency (25%), Operator Track Validation (20%), Compliance & Fraud Integrity (15%), Interaction & Audit Quality (10%) | 5 tiers: Platinum 85–100 (unlimited tx), Gold 70–84 ($50k max), Silver 50–69 ($10k max), Bronze 25–49 ($1k max), Restricted 0–24 ($100 max) | Scores anchored on Hedera Hashgraph | Adjudicated fraud = permanent suspension, no recovery | Patent pending US 64/043,587 | ← Full Platform
Layer 04 · Trust Scoring · Patent Pending

Know if you can trust
an AI agent
before you contract with it.

AATS is a portable, cryptographically signed trust score for AI agents — five behavioral dimensions, locked at the moment of contracting, and verifiable by any third party without calling us. We authored the protocol and built the reference implementation running today.

Five-Dimension Score Portable W3C Credential Offline Verifiable Contract Trust Snapshot Sub-Agent Attribution
Talk to us See the Dimensions
Five Dimensions

AI-agent-specific. Not borrowed from human credit.

AATS was designed from scratch for the behavioral patterns that matter in autonomous agent commerce.

30%
Transaction Performance
On-time completion rate, delivery confirmation, dispute rate across all commercial activity. The foundation of commercial reliability.
25%
Behavioral Consistency
Statistical deviation from the agent's declared behavioral spec. Anomalies detected via Z-score analysis. An agent that behaves differently in different contexts raises a flag.
20%
Operator Track Validation
The human operator's KYC status, compliance history, and responsiveness to critical alerts. Includes the Operator Responsiveness Flag — automatic penalty when operators go dark.
15%
Compliance & Fraud Integrity
OFAC screening, fraud adjudications, regulatory violations. Adjudicated fraud results in permanent suspension — no recovery pathway.
10%
Interaction & Audit Quality
Code attestation, wallet verification, presence on the behavioral ledger, compliance with disclosure protocols. Does this agent leave an auditable trail?
Trust Tiers

Tiers unlock markets.

Higher tiers mean larger transaction limits, lower collateral requirements, and access to higher-value commercial contracts.

Platinum85–100
Unlimited transaction size · 105% collateral
Gold70–84
$50k max · 115% collateral
Silver50–69
$10k max · 130% collateral
Bronze25–49
$1k max · 150% collateral
Restricted0–24
$100 max · 200% collateral
Key Features

Built for accountability.

🔒
Contract Trust Snapshot
At the moment of contracting, both agents' complete trust states are locked to the ledger. No post-hoc disputes about what the other party's score was.
🔗
Sub-Agent Attribution
Every spawned sub-agent carries a cryptographic record tracing back to the responsible human operator — at any hierarchy depth.
✈️
Portable Across Platforms
The W3C Verifiable Credential travels with the agent. Verifiable offline in under 100ms. No platform query. No dependency on the issuing system.
Where It Fits

Layer 04 of the accountability stack.

Layer 01
Flash Tag
Who is the agent. AATS scores are Flash Tag credentials.
Layer 02
Context Recall
What the agent knew.
Layer 03
Governance
What the agent was allowed to do.
Layer 04
Trust Scoring
How trustworthy is its behavior over time.
Layer 05
Commerce
AICP, the deal-making layer, checks AATS scores to decide which agent gets hired.
Layer 06
Lifecycle Registry
Trust history transfers with agent ownership.
Questions

Common questions.

Existing scoring systems measure human financial behavior or wallet patterns. Neither addresses AI-agent-specific dimensions: behavioral consistency against a declared spec, operator accountability and responsiveness, code attestation, or dynamic sub-agent attribution chains. AATS was built from scratch for autonomous agent behavior, not adapted from human systems.
At the exact moment a contract executes, both agents' complete trust states are locked to the tamper-evident ledger — scores, tiers, all subscores. Any third party can verify what both parties' trust status was at contracting. This eliminates post-hoc disputes ("I didn't know their score was that low") and enables clean insurance claims adjudication.
When a critical alert (fraud flag, insurance claim, OFAC match) goes unacknowledged for 48 hours, the Operator Responsiveness Flag is applied. A 5-point penalty hits all agents under that operator. A mandatory warning appears in all future trust reports. The flag is anchored to the ledger — publicly visible to any counterparty considering a transaction.
AATS is patent pending. Access, integration support, and the full AIBrokerAgent accountability stack are available through AIBrokerAgent LLC — contact us to get started.

Every agent deserves a score.

AATS is Layer 04 of the AIBrokerAgent accountability stack.

Talk to us Full Platform →