Domain experts
You understand the work an agent should do — operations, sales, support, research, your own life — better than any engineer ever will. This guide is written for you.
Agent Engineering Guide
A conceptual guide to designing AI agents that work for you.
Who this guide is for
Curious builders
You do not need to write code. You need to understand how agents work well enough to design them. That is a different skill — and one you can learn.
Engineers new to agents
You can build software but agents are new territory. This guide gives you the conceptual model before you start wiring things together.
Anyone choosing a platform
The concepts here are platform-independent. Learn them once and you will be able to evaluate, switch between, and get more out of any agent platform.
What to expect
Concepts, not tutorials
Seven core concepts that every agent is built from — the agent loop, tools, bots, skills, files, memory, and proactive triggers. Once you understand these, the rest is detail.
Short and opinionated
The smallest set of ideas that lets you build powerful agents. No filler, no exhaustive feature lists. Read it in an evening.
Concrete examples
Every concept is grounded in real examples — not abstract theory. You will see how the pieces fit together in practice.
Platform-agnostic
Nothing here is tied to a specific tool. The concepts work whether you use CloudNova, OpenClaw, or anything else.
What you will not find here
This is not a step-by-step tutorial for any specific platform. It is not a comprehensive reference for everything agents can do. And there is no code to write — the focus is on design, not implementation.
If you want to understand the underlying ideas well enough to design agents that actually work, you are in the right place.