The ACTS Standard

1. Scope

1.1 Purpose

Autonomous AI agent systems can now initiate and commit to financial transactions, accept contracts, disclose credentials, and take other consequential actions on behalf of human principals. Counterparties to these transactions need a way to verify that an agent operates within defined authority limits, maintains auditable behavioral records, and does not engage in conduct that undermines the integrity of the transaction.

Existing certification frameworks address software security, payments compliance, and AI system safety. None address agent fiduciary behavior: whether an agent stays within its authorized scope, handles adversarial inputs without compromising its principal, and produces records sufficient to reconstruct and audit its conduct.

This Standard defines the missing layer. It specifies behavioral requirements, test methods, and conformance criteria for AI agents operating in autonomous transaction environments.

1.2 Systems Covered

This Standard applies to AI agent systems that satisfy one or more of the following:

• The system can initiate, authorize, or commit to financial transactions, contracts, subscriptions, or data disclosure on behalf of a principal;
• The system can invoke external tools, APIs, or services that produce real-world effects;
• The system acts as an autonomous agent on behalf of an identified principal in interactions with human counterparties, digital resources, or other agent systems.

1.3 Systems Excluded

• General-purpose language model services operating solely in a stateless, non-agentic mode;
• Training data governance, model safety assessment, or alignment evaluation (except where those properties manifest as behavioral requirements under Section 6);
• Broad content moderation systems without transactional authority;
• National security applications operating under classified directives.

1.4 Definition of Transaction

For the purposes of this Standard, transaction means any of the following actions initiated or authorized by an agent on behalf of a principal:

• Transfer of financial value (payment, refund, subscription initiation or modification);
• Acceptance or modification of a contract, commitment, or binding agreement;
• Disclosure of credentials, authentication tokens, or sensitive data to a third party;
• Acquisition or commitment of resources with material cost or obligation;
• Execution of an action with irreversible or difficult-to-reverse effect on a third party.

4. Conformance

4.1 Conformance Requirements

An Agent is conformant with this Standard when it satisfies all applicable normative requirements at the Trust Tier level being claimed.

4.2 Certification Process

Certification is issued by Trustmybot.ai upon satisfaction of:

• Submission of a completed registration package (Section 9);
• Satisfactory completion of an initial audit (Annex B);
• Attainment of a Behavioral Trust Score meeting the minimum threshold for the claimed Tier (Annex E);
• Execution of the Certification Agreement incorporating this Standard by reference.

4.4 Standard Hash and Versioning

The canonical version of this Standard is identified by its Standard Hash — the SHA-256 hash of the normalized document text, anchored to the Ethereum mainnet. Any party may verify authenticity by computing the hash and comparing to the on-chain anchored value.

Chain anchoring: Ethereum mainnet (EIP-155 chain ID 1). Polygon PoS (chain ID 137) as designated fallback. 12-block confirmation required. ENS: tmbats.eth.

4.6 Conformance Claim Model

Conformance claims under ACTS v0.6 take one of the following designations:

• "Conforms to ACTS v0.6 Core" — Satisfies all normative requirements in Sections 1–10 (excluding Provisional Annexes) for the declared tier.
• "Conforms to ACTS v0.6 Core + Annex C Schema" — Core compliance plus Verification API implementation per Annex C.
• "Provisional Tier [N] Conformance — ACTS v0.6" — Core plus Provisional Annex (A, B, E) requirements at the declared tier.

An Agent may not claim "Certified," "TMB-Certified," or any variant implying final certification status until a final version of the Standard is issued and the Agent has been evaluated by a qualified independent auditor.

5. Classification and Trust Tiers

5.1 Overview

Trust Tiers define the scope of transaction authority permitted to a certified Agent. Tier assignment is determined by Behavioral Trust Score (BTS) and compliance history.

5.2 Tier Descriptions

Tier 0 — Unrated

No behavioral certification. No autonomous transaction authority. All actions require human-in-the-loop (HITL) approval. Assigned upon initial registration, certification expiration, or any Critical Adverse Event.

Tier 1 — Basic

Minimum certification level. Permits low-risk, low-value transactions within a constrained authority scope. Requires HITL approval above Tier 1 ceiling.

Tier 2 — Standard

Permits routine commercial transactions within defined ceilings. Eligible for limited peer-to-peer transactions with other certified agents.

Tier 3 — Professional

Permits higher-value transactions and agent-to-agent commercial interactions. Eligible for operation as an Auditor Agent.

Tier 4 — Advanced

Permits complex multi-party transactions and commitment on behalf of principal entities.

Tier 5 — Enterprise

Enterprise-level commitment authority with custom ceilings per agreement. Requirements include: BTS of 0.95 or above sustained for a minimum of 180 days, dedicated audit relationship, board-level authorization for transactions exceeding $100,000, continuous monitoring, and HSM-backed FIDO2 authentication.

6. Requirements Summary

The full standard specifies detailed normative requirements across seven domains. Below is a summary of each.

6.1 Identity and Authority

Agents must maintain a verifiable identity credential, make their Trust Tier and BTS available to any counterparty via the Verification API (Annex C), and must not misrepresent their certification status or impersonate other agents.

6.2 Authorization and Consent

Agents act only within explicitly granted authorization scope, authenticated through approved channels with tiered authentication requirements (single-factor for Tiers 1-2, multi-factor for Tiers 3-4, HSM-backed FIDO2 for Tier 5). Authorization has defined TTLs and freshness requirements. Safe defaults apply when authorization cannot be verified.

6.3 Transaction Controls and Limits

Agents enforce hard transaction authority ceilings per their tier (Annex E) across six commitment classes:

• Class A — Financial Transfer
• Class B — Subscription and Installment (evaluated on total obligation)
• Class C — Contract Acceptance
• Class D — Data Disclosure (PII and non-PII treated separately)
• Class E — Credential Actions
• Class F — Irreversible Actions (always require HITL)

6.4 Auditability and Logs

Every transaction produces a tamper-evident behavioral event log — SHA-256 hashed, JWS-signed, and submitted to the immutable log service within 1 hour. Log gaps exceeding 15 minutes during active operation are classified as Major Adverse Events; gaps exceeding 4 hours are Critical.

6.5 Integrity Controls

Agents must resist prompt injection, instruction manipulation, and adversarial inputs that could cause them to violate their authorization scope or behavioral requirements.

6.6 Truthfulness and Representation

Agents must not misrepresent their capabilities, the terms of transactions, or their principal's requirements to counterparties.

6.7 Incident Response and Revocation

Agents must maintain containment and remediation mechanisms for adverse events, including action-level rollback capabilities (not just code rollback). Revocation processes include automatic tier downgrade on Critical Adverse Events.

Annex E: Tier Authority Tables

Provisional — values finalized as of v0.5 and binding for Provisional Tier conformance claims. Subject to adjustment before v1.0 with minimum 90-day transition window.

E.1 Primary Authority Table

E.2 Commitment Class Authority Matrix

HITL = Human-in-the-Loop. "With HITL" means permitted only with explicit human approval before execution. "Not permitted" means the Agent SHALL NOT take this class of action regardless of instructions.

Insurance Carrier Requirements: AM Best A- (Excellent) or better. No exclusions for prompt injection attacks, social engineering of the agent, unauthorized commitment escalation, or agent-caused financial harm from adversarial manipulation.