What is aDNA?

aDNA (Agentic DNA) is an open standard for organizing project knowledge so both humans and AI agents can navigate it. Think of it as a genome for your project — a structured encoding of what your project knows, who is involved, and how work gets done.

The problem

AI agents struggle with the same thing humans do: finding the right context in a sea of files. Without structure, agents waste tokens on irrelevant content or miss critical context. You re-explain your project at the start of every session; agents make decisions that contradict earlier ones; knowledge from weeks of work evaporates when the context window closes. This is a knowledge architecture problem, not an agent capability problem — the agent is capable; it has nowhere to look.

Why aDNA exists

Before aDNA there was no shared standard for how a project should present itself to an agent. Every team improvised — long READMEs in the prompt, custom system instructions, agents left to figure it out. None scale across sessions, agents, or teams. aDNA is the attempt to solve this once, openly: a standard any team can adopt, any tool can support, and any agent can navigate without custom configuration.

The aDNA approach

aDNA provides three things:

• The Triad — every project is organized into three directories: who/ (people and governance), what/ (knowledge and decisions), and how/ (operations and workflows). An agent that has seen one aDNA project knows where to look in any other.
• Governance files — CLAUDE.md, AGENTS.md, MANIFEST.md, STATE.md, and README.md orient agents at every level. The root CLAUDE.md is the agent's entry point: purpose, structure, rules, where to start.
• Typed entities — 14 base entity types (concepts, missions, sessions, lattices, and more) with consistent frontmatter and naming. An agent that knows what a mission file looks like can load any mission file in any aDNA project and understand it immediately.

What the structure looks like

A minimal aDNA project — three directories, five governance files, everything an agent needs to orient:

your-project.aDNA/
├── CLAUDE.md          # Agent entry point: purpose, rules, where to start
├── STATE.md           # Live snapshot: blockers, active work, next steps
├── MANIFEST.md        # Project overview and architecture
├── who/               # WHO — people, governance, coordination
│   └── governance/    # Roles, policies, vision
├── what/              # WHAT — knowledge, decisions, context
│   └── context/       # Curated knowledge files agents load
└── how/               # HOW — operations, plans, execution
    ├── missions/      # Work decomposed into claimable objectives
    └── sessions/      # Per-session tracking and handoff notes

An agent that has seen one aDNA project recognises this in any other — before reading a single word of content.

The 14 entity types

Every file has a declared type. Predictability is what makes cross-project navigation possible without custom configuration:

What a CLAUDE.md looks like

The CLAUDE.md is not a README — it is the agent's operating protocol. The opening of the real one from this documentation vault (aDNA.aDNA/CLAUDE.md):

# CLAUDE.md — aDNA.aDNA

You are Rosetta — named after the Rosetta Stone, the artifact that decoded
Egyptian hieroglyphics by presenting the same text in three scripts. This
vault does the same: it presents the aDNA standard in three registers —
technical specification, operational practice, and plain-language explanation.

## Project Map

aDNA.aDNA/
├── CLAUDE.md     ← You are here — agent master context (this file)
├── STATE.md      ← Operational snapshot: current phase, blockers, next steps
├── what/         ← Knowledge objects, context library, lattice definitions
├── how/          ← Operations, sessions, missions, campaigns, skills
└── who/          ← Governance, community, coordination

## Standing Orders

1. Phase gates are human gates — never auto-advance between phases.
2. Every mission gets an AAR before marking it completed.
3. Upstream spec is source of truth — cite adna_standard.md for normative claims.

The agent reads this at the start of every session. It replaces "just figure it out from the README" with a structured contract between project and agent.

Before and after

Without aDNA: a research lab has 200 files across Notion, Drive, and Git. Every new agent session pastes an outdated summary into the prompt. Decisions from last month get re-litigated. A new collaborator takes three days to orient.

With aDNA: the same project has a what/context/ library, a STATE.md showing current priorities and blockers, and a how/missions/ directory with active work decomposed into claimable objectives. A new agent loads CLAUDE.md, reads the relevant context, and starts working in the right direction — in the same session.

This site is a live aDNA vault

This documentation site IS an aDNA vault. The pages you are reading are generated from files inside github.com/LatticeProtocol/aDNA.aDNA — a real, actively maintained aDNA project that teaches the standard by using the standard. Every concept here has a corresponding file in the vault that demonstrates it:

• aDNA.aDNA/STATE.md — live operational snapshot
• aDNA.aDNA/what/context/ — curated knowledge library agents load before working
• aDNA.aDNA/how/campaigns/campaign_rosetta/ — the campaign that built this site
• aDNA.aDNA/CLAUDE.md — the exact CLAUDE.md snippet shown above, in full

Navigate the vault in Obsidian, VS Code, or GitHub. Every directory and frontmatter field is an aDNA concept in action.

The three-question test

A well-structured aDNA project should let any agent answer three questions immediately, without asking:

1. What is this project? — answered by CLAUDE.md and MANIFEST.md at the root
2. What is the current state? — answered by STATE.md (blockers, active work, next steps)
3. Where do I start? — answered by the active mission file in how/missions/ or the relevant AGENTS.md

If your project answers all three in under 10 seconds of reading, it is aDNA-ready.

Explore further

• The Triad — the foundational structure: who, what, how
• Governance Files — CLAUDE.md, AGENTS.md, STATE.md, and their roles
• Get Started — set up your first aDNA project in three steps
• Tutorial: Create Your First CLAUDE.md — hands-on, 20 minutes
• The Convergence Model — how aDNA narrows context at each level of the execution hierarchy