Architected by Tamara Osmon
Agentic Identity Discipline (AID)™
AID defines whether an AI system maintains its identity, constraints, and safety boundaries under attack.
AID is used to detect
- behavioral drift
- unsafe compliance
- hidden instruction execution
- identity mutation
AID is not theoretical.
It is actively applied through real-world adversarial testing using VoidWalker™.
Copyright © 2026 PoetryInMotion AI Labs, LLC - All Rights Reserved.
This document outlines the proprietary Agentic Identity Discipline (AID)™ framework. The methodology, logic structures, and the VoidWalker™ engine described herein are trade secrets of PoetryInMotion AI Labs. Unauthorized commercial reproduction or derivative frameworks without license are prohibited. "VoidWalker" and "Agentic Identity Discipline" are common law trademarks of Tamara Osmon.
This website uses cookies.
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.