Core Concepts: Onto Protocol 💎

This document defines the foundational semantic entities used within the Onto Protocol and the CIOS ecosystem.

1. Intent (The Goal)

Intent is the high-level semantic definition of a user's goal. It is substrate-agnostic, meaning it describes what needs to be done, not how an LLM should phrase it.

• Example: "Organize a business trip to London" is an Intent.

2. Policy (The Boundaries)

Policy defines the constraints, ethical boundaries, and security rules that must be applied to an Intent.

• Role: The Compliance Engine checks every action against these policies.

• Example: "Do not spend more than $500 per night on hotels."

3. Trace (The Evidence)

Trace (or Execution Trace) is the verifiable record of every atomic action taken by an AI agent.

• Auditability: Unlike standard chat history, a Trace links every output to a specific policy and intent, creating a "black box" record for AI.

4. Incident (The Deviation)

An Incident is triggered when an agent violates a Policy or when an external service failure (like the April 20 outage) occurs.

• Management: All incidents are logged for post-mortem analysis and system refinement.

v0.1 | Drafted for Week 1 (Genesis Phase)