ΔOS terminology

A structured representation of what an agent wants to do. Includes action type, target, parameters, and context.

Intent Intermediate Representation. The canonical schema for expressing agent intents within ΔOS.

The decision rendered by ΔOS for a given intent: allow, allow with conditions, escalate, or block.

Judgment Intermediate Representation. The canonical schema for governance decisions.

A declarative rule that defines governance requirements for specific types of actions or contexts.

Policy Intermediate Representation. The canonical schema for expressing governance policies.

The result of executing an action: success, failure, or partial completion.

Outcome Intermediate Representation. The canonical schema for recording action results.

A classification of action risk from T0 (read-only) to T3 (critical). Higher tiers require more governance.

A domain-specific model that understands agent actions in context. Classifies intents and assesses risk.

The core decision engine that receives intents, evaluates policies, and renders judgments.

A preset configuration that defines governance posture: Regulated Enterprise, Enterprise Core, Digital Native, or High Velocity.

A record of business value delivered by governance decisions: time saved, risk avoided, cost prevented.

A complete record of a governance decision including intent, policy evaluation, judgment, and outcome.

An isolated organizational unit within ΔOS. All data and policies are tenant-scoped.

A deployment context (dev, staging, production) that may have different governance requirements.

A functional area (SDLC, Security, Infrastructure) with domain-specific governance patterns.

Routing a decision to a human reviewer when policy requires approval.

Reversing an action that has already been executed. Available for actions classified as reversible.

The scope of potential impact from an action. Used to classify risk and determine governance requirements.