Skip to the content.

DDAD Artifact Model

Overview

DDAD is implemented using a small, explicit set of repository artifacts that govern AI execution.

The Artifact Flow

README.md
  ↓ explains purpose

AGENTS.md
  ↓ governs AI behavior

TODO.md
  ↓ authorizes work

LLD Markdown
  ↓ defines implementation

Code + Tests

Artifact Descriptions

README.md

Purpose: Explains the project purpose and context

Contains:

Role: Provides context for all other artifacts

AGENTS.md

Purpose: Governs AI behavior and constraints

Contains:

Role: Defines what AI can and cannot do

TODO.md

Purpose: Authorizes specific work items

Contains:

Role: Authorizes AI to work on specific tasks

LLD Markdown

Purpose: Defines exact implementation requirements

Contains:

Role: Specifies exactly how to implement

Artifact Relationships

README.md (context)
    ↓
AGENTS.md (governance)
    ↓
TODO.md (authorization)
    ↓ references
LLD Markdown (specification)
    ↓
Code + Tests (implementation)

Key Principles

1. Explicit Over Implicit

2. Version Controlled

3. Human Authored

4. Single Source of Truth

Artifact Quality

Good Artifacts

Poor Artifacts

Best Practices

1. Start with README.md

2. Define AGENTS.md Early

3. Create LLD Before TODO

4. Keep Artifacts Synchronized