AI & Agents
This section contains templates and guidelines for working with AI assistants, building agentic systems, and establishing effective human-AI collaboration patterns.
Why This Matters
In an agentic world, AI systems need context to be effective. Well-structured documentation helps AI assistants:
- Understand your codebase, conventions, and constraints
- Make better suggestions aligned with your architecture
- Avoid common pitfalls specific to your project
- Execute multi-step tasks autonomously and safely
Documentation Categories
| Category | Purpose | Key Artifacts |
|---|---|---|
| Context Files | Give AI agents understanding of your codebase | AGENTS.md, CLAUDE.md |
| Agent Design | Design AI-powered features and autonomous agents | ADD (Agent Design Docs) |
| Prompts | Reusable prompt templates for common tasks | System prompts, task prompts |
Philosophy
“Context is the new code” — The quality of AI assistance is directly proportional to the quality of context you provide.
How to Use These Docs
- Starting a new repo? Create an AGENTS.md to onboard AI assistants
- Building an AI feature? Write an Agent Design Doc to define capabilities and guardrails
- Repeating prompts? Capture them in Prompt Templates for consistency
- Reviewing AI output? Use the evaluation criteria from your ADD