Governed Autonomy Architecture

The reference architecture: five Planes that place responsibilities. They do not prescribe a product topology.

The Architecture places where governing functions sit and what each is responsible for. It is part of the v4 text; the full statement, with the mandatory runtime baseline, is on the Doctrine v4 page.

The 5 Architectural Planes

Plane 1: Agent Identity & Lifecycle

Attribute actors and workloads; bind approved missions and grants; manage issuance, change, expiry and revocation.

Plane 2: Execution & Tool Governance

Mediate effects, enforce boundaries, bind evaluated work to execution, and govern queues, retries and indirect paths.

Plane 3: Policy & Compliance Engine

Evaluate current mission context, authoritative relationships, shared limits, obligations and policy composition.

Plane 4: Human Oversight, Audit & Traceability

Provide intervention and accountable continuity; preserve sufficient, protected evidence and expose uncertainty.

Plane 5: Multi-Agent Trust & Delegation

Establish authority at each handoff; preserve grant chains, obligations, cumulative constraints and affected-output lineage.

Enforcement and decision functions

Planes 2 and 3 distinguish the policy enforcement point from the policy decision function. The agent shall not be able to replace or bypass the functions that govern it. Logical separation can be implemented in several ways, but the claimed separation must survive the actor and paths within the assessed boundary.

Vocabulary note

An unqualified Plane means one of the five Architectural Planes. Other established senses, such as a network control plane, are qualified. A runtime harness is the execution and tool-governance component, not a sixth Plane or a replacement doctrine name. The baseline, assessment companion and organizational maturity evaluation are separate instruments, not additional Architectural Planes.