# AI Harness Doctrine

> The architectural doctrine for runtime governance of autonomous AI agents across enterprise systems. Authorize the Agent. Govern the Behavior. Published by the AI Harness Institute. Current version: v3.1.

AI Harness defines a new enterprise control discipline: governing autonomous AI agents as first-class identities through continuous runtime enforcement, across all systems they touch, at the level of behavior — not just access. It is to autonomous AI what Zero Trust was to network security. Core vocabulary: the 5 Laws, the 5 Architectural Planes, the 6 Framework Pillars, Least Agency, the Threat Surface (prompt injection, intent hijacking, cascading failure, behavioral drift), and a three-level Maturity Model.

## The Doctrine

- [Declaration](https://aiharnessdoctrine.org/): the case for behavioral governance of autonomous agents — the gap, the threat surface, the three non-negotiables
- [The 5 Laws](https://aiharnessdoctrine.org/laws/): Agents Are Identities, Not Tools · Enforce at Runtime · Governance Must Span Systems · Trust Does Not Travel · Humans Retain the Right to Intervene
- [Architecture](https://aiharnessdoctrine.org/architecture/): the 5 Architectural Planes — Agent Identity & Lifecycle, Execution & Tool Governance, Policy & Compliance Engine, Human Oversight Audit & Traceability, Multi-Agent Trust & Delegation
- [Framework](https://aiharnessdoctrine.org/framework/): the 6 Pillars — the implementation methodology from principle to practice
- [The Zero Trust Parallel](https://aiharnessdoctrine.org/zero-trust/): why this is the next architectural shift
- [Maturity Model](https://aiharnessdoctrine.org/maturity/): Level 1 Identified → Level 2 Governed → Level 3 Continuous

## Comparisons

- [AI Harness vs the Existing Stack](https://aiharnessdoctrine.org/vs/): where IAM, security monitoring, orchestration, and AI guardrails each stop
- [vs Identity & Access Management](https://aiharnessdoctrine.org/vs/iam/)
- [vs Security Monitoring](https://aiharnessdoctrine.org/vs/siem/)
- [vs Orchestration & Workflow Automation](https://aiharnessdoctrine.org/vs/orchestration/)
- [vs AI Guardrails](https://aiharnessdoctrine.org/vs/guardrails/)

## Meta

- [Changelog](https://aiharnessdoctrine.org/changelog/): the doctrine is versioned, dated, and revised in the open