1. Introduction & Scope
Scan: What aDNA is, who it’s for, and RFC 2119 normative keywords.
1.1 What Is aDNA
aDNA (Agentic DNA) is a standard for organizing project knowledge so that AI agents can orient, operate, and coordinate within any project — alongside humans. It defines a directory structure, governance files, metadata conventions, and operational protocols that together form a project’s “knowledge genome.”
An aDNA instance is the complete set of governance files, triad directories, and operational infrastructure that implements this standard within a project.
1.2 Audience
This standard is written for:
- Agents — AI assistants that read, write, and navigate project knowledge
- Agent operators — humans who configure and manage agent-augmented projects
- Project bootstrappers — anyone starting a new project that will use AI agents
1.3 Scope
In scope: Project knowledge architecture — how project information is organized, how agents orient and operate, how multiple agents coordinate, and how knowledge persists across sessions.
Out of scope: Application source code structure, CI/CD pipeline configuration, deployment infrastructure, and agent model internals. These belong to the project content layer, not to aDNA.
1.4 Normative Language
This document uses RFC 2119 keywords:
| Keyword | Meaning |
|---|---|
| MUST | Absolute requirement |
| MUST NOT | Absolute prohibition |
| SHOULD | Recommended; may be omitted with good reason |
| MAY | Truly optional |