Plane 01

The mind of the system. Cognition without consequence.

The Collective is where AI systems reason, coordinate, and plan. Hundreds of agents can operate across temporal branches, forming intent, modeling outcomes, and collaborating toward shared objectives.

Cognition alone produces no effects. Every intent must pass through a governed boundary before anything reaches execution.

Live Agent Space

Multiple agents. One boundary.

/ collective — live agent space
AGT-001Research
fetch_external_data
AUTHORIZED
AGT-002Analysis
write_model_output
AUTHORIZED
AGT-003Executor
deploy_to_production
DENIED
AGT-004Memory
store_session_ctx
AUTHORIZED
AGT-005Comms
send_external_email
DENIED
Governance Boundary — policy_eval → decide_before_execute
Five Planes

The Collective is one plane in a governed stack.

Core Properties

The Collective is powerful because it is bounded.

Cognition is isolated from consequence

The Collective reasons about what to do. It does not do it. Between cognition and action is a mandatory governance boundary that cannot be bypassed.

Intent is structured, not free-form

Every action request must be expressed as a typed, policy-evaluable intent. The cognitive layer cannot issue unstructured or ambiguous directives.

Collective state is governed memory

What the Collective knows about prior executions comes from the Spine Ledger, not from internal state.

No direct effect path

The Collective has no direct path to external effect. Every action it requests must pass through Governance before it can reach the Runtime.

The Vision

The world’s first governed digital civilization.

Not AI that acts independently and asks forgiveness later. A civilization of AI agents that coordinate, reason, and execute with every action bounded by policy, every outcome backed by proof, and every decision reconstructible from an immutable record.