Agentic Software Engineering

Need a Software? Let's "ENGINEER IT!" A Guide to Building Verified, Secure, and Operable Software with AI Agents and Harnesses

A controlled lifecycle for building software with AI agents, from need and specification through tests, release and operation.

Who it is for. Software practitioners and technically curious beginners who want to use AI coding agents through a checkable engineering lifecycle.

It is being checked before release. Ask, and we will tell you when it is ready.

What it covers

Teaches a controlled lifecycle for using AI coding agents, taking one need from exploration and specification through grounding, design, test-driven implementation, evaluation, release, operation and retirement. Each chapter gives a method, a worked demonstration, a practice activity and a reusable record.

15 chapters in 3 parts

Part IAI Foundations for Software Engineering

Chapter 1: AI Literacy for Software Engineering
Chapter 2: Software Engineering from Need to Retirement
Chapter 3: Agentic Engineering Systems and Human Responsibility

Part IIThe ENGINEER IT Process

Chapter 4: Explore the Need and System (E)
Chapter 5: Name the Outcome and Acceptance Evidence (N)
Chapter 6: Ground the Work in Reality (G)
Chapter 7: Imagine and Narrow the Solution (IN)
Chapter 8: Execute Through Specification-and-Test-Driven Slices (E)
Chapter 9: Evaluate and Recover (ER)
Chapter 10: Integrate and Release (I)
Chapter 11: Tend the System Through Its Life (T)

Part IIIApplied Agentic Engineering

Chapter 12: Capstone: Engineer a Verified Working Application
Chapter 13: ENGINEER IT for Existing Systems and Repair Work
Chapter 14: ENGINEER IT for Teams and Multiple Agents
Chapter 15: Governing the Agent Through Repository Instructions: Design, Empirical Limits, and Continuous Auditing of AGENTS.md