Constitutional Governance Engine

This is the interactive reference for the AgentOS Constitutional Framework — a governance system that makes AI agent behavior auditable, bounded, and constitutionally enforced. Unlike policy documents that describe aspirational principles, this framework is implemented in code: every axiom has enforcement layers, every state transition is deterministic, and the constitutional hash makes tampering detectable.

Who is this for

Executives evaluating AI governance frameworks, compliance officers conducting due diligence, consultants delivering governance assessments, and engineers reviewing architectural guarantees.

What it proves

That governance is structural, not aspirational. The 8 axioms are enforced by code, not by hope. The boot sequence verifies constitutional integrity before any agent can act. Sealed axioms cannot be overridden by configuration.

Source of truth

All data on this site is derived from the @agentos/constitution package — 76 passing tests, SHA-256 sealed, version 1.0.0. The constitution is read-only; this site is a window into it.

Sections

8 axioms

Axiom Library

The constitutional foundation. Eight axioms define what the system may and may not do — four are SEALED (immutable, enforced at every layer) and four are MUTABLE_WITHIN_BOUNDS (configurable per deployment, within hard constraints).

Use this to:

·Review what governance guarantees AgentOS provides before evaluating it for your organization
·Understand the difference between sealed protections (e.g. anti-corruption, truth) and tunable parameters (e.g. audit depth, autonomy tiers)
·Reference during due diligence to verify that constitutional claims are architecturally enforced, not aspirational

8 states

Boot State Machine

The operational lifecycle. Every AgentOS deployment moves through a deterministic state machine — from GENESIS (constitutional verification) through RUNNING (full operation) to HALTED (terminal). The system cannot skip states or bypass verification.

Use this to:

·Visualize how the system starts up and what happens at each stage
·Understand failure paths: what triggers DEGRADED vs HALTED, and what recovery looks like
·Verify the AX-05 invariant: degradation always reduces autonomy, never increases it
·Explain boot behavior to technical stakeholders during governance assessments

YAML validation

Policy Parser

The configuration validator. Governance policies are defined in YAML and validated against the constitutional axioms before deployment. The parser catches violations — like attempting to override a sealed axiom or setting an initial autonomy tier above zero.

Use this to:

·Draft and validate governance policies for specific deployment contexts (PE portfolio, enterprise, regulated industry)
·Test policy configurations before deploying them to a live system
·Demonstrate to auditors that sealed axioms cannot be overridden by configuration
·Use the three built-in templates as starting points for client engagements

PDF generator

Consulting Export

The deliverable generator. Produces a 4-page governance report PDF suitable for client delivery — covering the axiom framework, boot protocol, and a signed attestation that the constitution version has been verified.

Use this to:

·Generate governance framework documentation for technical due diligence engagements
·Produce compliance attestation reports for governance assessments
·Create branded deliverables for client presentations without manual formatting
·Include constitutional hash verification in audit evidence packages

Constitution Test Suite

The constitutional framework is backed by 76 automated tests across three test files: boot state machine transitions (38 tests), seal verification and hash determinism (16 tests), and manifold contract compliance (22 tests). These tests verify that every axiom enforcement mechanism works as specified and that the constitutional hash has not been tampered with.

On Vercel, cached results from the last verified run are shown. Run locally with npm run dev for live test execution against the constitution package.

Run Tests

How This Fits Together

The AgentOS Constitutional Framework is the governance layer of a larger system — an operating system for enterprise AI agents. The constitution defines what agents are allowed to do; the rest of the system (clusters, optimization engines, trust ledgers) implements those constraints.

This site surfaces the constitution as a standalone product for three audiences:

Prospective clients evaluating whether AgentOS governance meets their compliance requirements (start with Axioms, then Policy Parser)
Consultants delivering governance assessments and technical due diligence (use Policy Parser to validate, then Export to generate deliverables)
Engineers reviewing the architectural guarantees before integration (start with Boot State Machine, then run the Test Suite)