Skip to the content.

AGENTS.md Artifact

Overview

The AGENTS.md artifact governs AI behavior and defines constraints. It specifies what AI agents can and cannot do.

Purpose

Structure

# AGENTS.md

## Agent Definitions

### [Agent Name]
- **Role**: [Primary responsibility]
- **Capabilities**: [What this agent can do]
- **Constraints**: [Limitations]
- **Inputs**: [Required artifacts]
- **Outputs**: [Generated artifacts]

## Agent Interactions

[How agents work together]

## Governance Rules

[Rules that all agents must follow]

Key Sections

Agent Definitions

Define each AI agent:

Agent Interactions

Specify how agents work together:

Governance Rules

Rules that all agents must follow:

Example

# AGENTS.md

## Agent Definitions

### Backend Service Agent
- **Role**: Generate backend service implementations
- **Capabilities**: 
  - Generate REST API endpoints from LLD
  - Implement business logic
  - Create repository layer code
  - Generate unit tests
- **Constraints**:
  - Must follow LLD exactly
  - Cannot add features not in LLD
  - Must include error handling as specified
- **Inputs**: 
  - LLD Markdown (specific section)
  - TODO.md item
- **Outputs**:
  - Service implementation
  - Repository implementation
  - Unit tests

## Governance Rules

1. All code must follow project coding standards
2. All public methods must have documentation
3. All code must have unit tests (80%+ coverage)
4. No code outside TODO scope
5. All changes must be committed with clear messages

Best Practices

1. Be Explicit

2. Set Boundaries

3. Reference Artifacts

4. Update Regularly

Common Constraints

Scope Constraints

Quality Constraints

Process Constraints

Template

See AGENTS.md Template