Developer Preview · In Development

Agent Reputation Management

Autheo is developing a reputation and risk scoring subsystem for agents, merchants, controllers, and missions. It is designed to help counterparties evaluate trust before accepting an autonomous action.

What Reputation Management is

Reputation Management is designed to attach a verifiable risk and trust score to each counterparty an agent encounters so the agent, its controller, or a verifying merchant can make an informed decision before authorizing an action.

• Scores agents, merchants, controllers, and individual missions.
• Anchored on KYA credentials so scores are bound to a verifiable identity rather than a free form handle.
• Designed to be queried at decision time by a route planner, a merchant verifier, or another agent.

Where it sits in the stack

Reputation Management spans multiple layers because trust is consumed at the identity layer, surfaced at the discovery layer, routed through the protocol abstraction layer, and informed by settlement outcomes.

• L7 Trust and Identity: scores are bound to KYA agent, controller, and merchant credentials.
• L5 Merchant and Catalog Discovery: merchant reputation can be surfaced alongside the merchant manifest.
• L1 OS and Protocol Abstraction: the route planner can consume reputation signals when choosing a protocol path.
• L3 Settlement and Clearing (optional): settlement outcomes and disputes feed back into the score as optional inputs.

Supported gaps

Reputation Management supports three of the five strategic gaps. It is not a new gap.

• KYA: extends KYA credentials with a trust evaluation layer.
• Protocol Router: gives the router a trust signal when comparing candidate routes.
• Long-tail Merchant Discovery: gives merchants outside large platform networks a verifiable way to surface trust.

Potential inputs

Inputs may be aggregated from on chain events, KYA credential status lists, mandate logs, merchant manifests, and counterparty confirmations.

• Completed missions and their outcomes.
• Failed missions and the reason for failure.
• Revoked credentials and KYA status list updates.
• Disputed transactions and dispute outcomes.
• Merchant feedback aggregated across missions.
• Settlement history across stablecoin, card networks, x402, and chain rails.
• Policy violations recorded against a mandate.
• Credential age and continuity.
• Controller reputation aggregated across the agents a controller operates.
• Compliance report availability for past missions.

Potential outputs

Outputs are designed to be machine readable and actionable at decision time.

• Reputation score for the agent, merchant, controller, or mission.
• Risk flags describing why the score is at its current level.
• Credential health summary including expiry, revocation, and continuity signals.
• Mission history summary across recent completed and failed missions.
• Counterparty attestations contributed by merchants and other agents.
• Dispute indicators for outstanding or recently resolved disputes.
• Route policy recommendations for the Protocol Router when reputation suggests preferring or avoiding a path.

What Reputation Management is not

Reputation Management is a technical trust signal for autonomous decisions and is not a substitute for human review.

• Not a public social score for end users.
• Not a governance voting mechanism. THEO remains a utility token.
• Not a replacement for compliance review by legal, regulatory, or accounting professionals.
• Not a credit score and not an investment rating.

Related documents