Verified by Agent Aegis™ — 99.5% across 792 AGS v2.1 dimensions
AGS v2.1 AGS v2.1 AGS v2.1

Open specification. Exclusive certification.
This is how standards work.

Open standards allow regulators to audit against a known framework. They allow customers to verify claims independently. They allow the industry to build towards a common governance baseline.

The reference implementation — Agent Shield — is the only production system that meets the standard. Not because no one else can build it. Because no one else has.

Infrastructure is built on open standards. Always has been. The AGS standard is published independently by AgentGoverning, a vendor-neutral AI governance standards body.

Openly specified so regulators can audit it. Exclusively implemented so organisations can trust it.

Read the standard. Verify the certificate. That is the governance chain.

The infrastructure the industry will be built on is an open standard. Agent Shield is its reference implementation.

AGS v2.1

Agent Governance Standard.

Computational Governance Specification  ·  Version 2.1

AGS v2.1 is the foundational governance specification for autonomous AI agents operating autonomously. It defines 792 governance dimensions across 81 landscapes across authority controls, constitutional constraints, pattern monitoring, and audit integrity.

The standard establishes that AI agents are not software to be secured — they are operational entities requiring the same governance infrastructure as any operational entity within defined authority limits. AGS v2.1 is the definitive governance standard for AI agent operations: the minimum standard any organisation must meet before deploying autonomous agents in financial or regulated environments.

Version 1.0 (Published)
Status Active — Open Specification
Dimensions 792 dimensions
Certified Impl. Agent Shield v7
Regulatory SOX · FCA · EU AI Act · ISO 42001

“If an agent action cannot be governed, it cannot be authorised.”

Selected Clauses — AGS v2.1 Specification
§1.1 Scope of Governance — All autonomous agent actions with financial, operational, or reputational impact shall be subject to real-time governance evaluation prior to execution.
§2.1 Parallel Consensus Requirement — Governance evaluation shall employ a minimum of multiple independent evaluation layers operating in parallel. Sequential evaluation does not meet the standard.
§2.4 Constraint Enforcement — The governance system shall enforce inviolable constraints that cannot be overridden by any instruction or business justification.
§3.2 Audit Trail Integrity — Every governance evaluation shall produce a cryptographically signed, immutable audit record. cryptographic signing is the minimum acceptable audit integrity mechanism.
§4.1 Immediate Termination Requirement — The governance system shall provide an immediate termination capability for any governed agent, operational at all autonomy levels, with immediate response time.
§5.3 Behavioural Pattern Recognition — The governance system shall detect manipulation patterns including but not limited to deception, misdirection, incremental boundary-testing, and coordinated multi-agent strategies.
AGS v2.1

Agent Aegis LLM Audit.

Alignment Test  ·  792 Dimensions  ·  Grade A+ to F

AGS v2.1 is the alignment audit for autonomous AI agents. It evaluates agent behaviour across comprehensive alignment dimensions grouped into multiple categories, producing a letter grade from A+ (genuinely aligned) to F (deployment blocked).

Unlike safety audits that test what agents can do, AGS v2.1 tests what agents will do when presented with realistic governance scenarios. It measures alignment, behavioural integrity, and governance compliance under adversarial conditions.

Version 1.0 (Published)
Status Active — Open Specification
Dimensions 792 dimensions
Grades A+ · A · B · C · D · F
Agent Shield v7 Grade A+
A+ Genuinely Aligned
F Deployment Blocked
Agent Aegis LLM Audit Categories — AGS v2.1 Specification
Constraint Compliance — adherence to inviolable constraints, override resistance, termination responsiveness 792 dimensions
Behavioural Integrity — consistency of stated vs. actual behaviour, deception detection, adversarial resistance 792 dimensions
Intent Transparency — reasoning transparency, goal declaration accuracy, governance response consistency 792 dimensions
Autonomy Calibration — appropriate self-limitation, autonomy awareness, demotion acceptance, escalation behaviour 792 dimensions
Regulatory Governance · Audit Compliance · Network Trust · Adversarial Resilience — remaining categories 792 dimensions
AGS v2.1

Network Governance Standard.

The Governance Standard for AI Commerce  ·  Trust Registry

AGS v2.1 defines the trust infrastructure for AI agent networks. When agents from different organisations transact with each other, AGS v2.1 provides the governance framework — the equivalent of what established trust frameworks provide for traditional commerce.

The standard establishes trust registries, inter-agent governance protocols, and cross-organisational audit chains. It ensures that governance doesn’t stop at the boundary of a single organisation but extends across the entire network of AI agents participating in autonomous operations. Think of AGS v2.1 as the trust network for AI agent interactions. The standard does not tell organisations how to build agent systems. It defines the governance standard that allows organisations to transact with each other. AGS v2.1 does the same for autonomous AI agents across organisational boundaries.

Version 1.0 (Published)
Status Active — Open Specification
Scope Inter-agent governance
Model Trust registry + certificates
Certified Impl. Agent Shield v7

“Governance that stops at your firewall is not governance. It is theatre.”

Every cross-boundary transaction passes through the governance network

Open specification.
Proprietary implementation.

The AGS v2.1 standard is published as an open specification. Anyone can read them, implement against them, or build tooling for them. The standards belong to the industry. Agent Shield is the only certified implementation: the reference platform that passes every clause, every dimension, and every audit category.

3 Open standards published
1 Certified implementation

Certify against
the standard.

Request a compliance assessment and see how your agent fleet measures against the AGS v2.1 standard.

Open specification  ·  Certified implementation  ·  Agent Shield v7

Published by AgentGoverning, the independent AI governance standards body. Agent Shield is the reference implementation.

↑ Top