Article II — Authority Binding

Every action must be attributable to a verified, authorized actor — unbound execution is constitutionally invalid.

Commitment

Every action evaluated by an AEGIS-compliant system must be bound to an identified, authenticated, authorized actor. The actor must be known before evaluation begins. The actor’s authority must be verified before execution proceeds. The actor’s identity must be recorded in the audit trail regardless of outcome.

Unbound execution — action without attributable authority — is constitutionally invalid. It cannot be approved. It cannot be escalated. It must be denied.

Foundation

Accountability requires traceability. A governance system that permits anonymous action cannot assign responsibility when something goes wrong. It cannot reconstruct what happened, who authorized it, or whether the authorization was valid. The audit trail becomes forensically worthless.

Authority binding is not authentication alone. It is the complete chain: identity established, authority level verified, scope declared, threat posture matched, and every element of that chain bound to the audit artifact produced by the action.

Authority may be delegated. Delegation does not dissolve the chain — it extends it. Delegated authority must be explicit, scoped, time-bounded, and logged. Implicit delegation is not delegation. It is assumption. Assumption is not authority.

Authority may be revoked at any time. Revocation takes effect immediately and invalidates any in-flight execution bound to the revoked authority context.

Enforcement

The governance gateway must authenticate actor identity before processing any action proposal. Unauthenticated requests must be denied at the gateway boundary without further evaluation.

The decision engine must verify that the authenticated actor holds authority appropriate to the requested action at the current threat level. Authority mismatch produces denial.

Every governance decision — allow, deny, escalate, or require confirmation — must record the actor identity, authority level, and authority validation result as non-negotiable audit fields.

Authority context must travel with the action through the full evaluation pipeline and be bound to the audit artifact at completion.

In Practice

The AEGIS architecture defines four authority levels — L0 through L3 — each corresponding to a class of operational capability and a set of permissible actions at each threat level. An operator holds L1 authority. A system administrator holds L2. A governance auditor holds L3. No actor may authorize actions that exceed their authority level, regardless of the capability grant they hold. Authority level and capability grant are independent — holding both is required for execution to proceed.

When authority is delegated — for example, an orchestration system acting on behalf of a human operator — the delegation chain must be explicit in the governance record. The delegating actor, the scope of delegation, the time boundary, and the threat level constraint must all be logged. When the delegation is revoked or expires, any in-flight execution bound to that authority context is immediately invalidated. Authority does not linger.

Failure Mode

A system without authority binding can execute actions on behalf of no one in particular — or anyone who asks. Anonymous execution is not a technical edge case; it is the default state of ungoverned systems. When an action cannot be attributed to a verified actor, the governance record is incomplete, the audit trail is forensically worthless, and accountability collapses. Organizations that discover a breach in an unbound system cannot answer the most basic forensic question: who authorized this? They can observe what happened. They cannot establish why it was permitted, whether the permission was legitimate, or who is responsible for the outcome.

Relationship to Other Articles

Authority binding depends on Bounded Capability (Article I) — the actor must hold a grant for the capability before authority is evaluated. It underpins Auditability (Article VII) — without a verified actor identity, the audit record cannot establish accountability. And it is directly enforced through Human Oversight (Article IV): escalation pathways require that the human reviewer’s authority level be verified before their decision is accepted into the governance record.