A field guide to professional AI-assisted engineering
Agentic Coding, done properly
Direct AI agents to ship production-ready software used by millions — with real tests, a staging pipeline, and safe releases. Not vibe coding: engineering, amplified.
AI can write code faster than you can read it. That is the opportunity and the danger. Point an agent (an AI that doesn't just chat but actually reads files, runs commands, and fixes its own errors in a loop) at a task with no structure, and you get a tangled mess you can't ship. Wrap it in the right workflow — plan, test, review, release — and the same agent ships features to production while you sleep. That directed way of working is agentic coding.
This guide teaches that workflow from zero. It is built on the methods of Matt Pocock (creator of the widely-used skills for real engineers) and the published engineering practice of Anthropic, Thoughtworks, GitHub, Simon Willison, Addy Osmani and others — all from 2025–2026.
What you'll be able to do
Take an idea, align an agent on exactly what to build, and have it write the feature test-first. Review it by watching it work. Ship it through a staging pipeline — behind a feature flag, with instant rollback — safely enough for a product real people depend on.
The whole method is one repeating loop. Six moves take an idea to a shipped, safe feature. The rest of this guide is these moves in detail, then the production, scaling and judgment layers around them.
Move 1 · §3
Align
Let the agent grill you until you both know exactly what to build.
Move 2 · §4
Plan
Turn that into a destination document and thin, testable slices.
Move 3 · §5
Build
Test-first, one test at a time — a separate agent writes the test.
Move 4 · §6
Verify
Watch a recorded demo of the behaviour instead of reading every line.
Move 5 · §7
Harden
Pass the quality gates: types, coverage, and the non-functionals.
Move 6 · §8
Ship
Through staging, behind a flag, with rollback one click away.
In a hurry?Jump to Setup → Align → Build. Come back for the production and scaling sections when you're ready to go live.
How to read this guide
New to software entirely? Section 0 explains what an agent is and the vocabulary this guide assumes — start there. Sections 2–6 are hands-on: you set up, then run the loop on a real task with copy-able commands and a "you'll know it worked" check at each step. Sections 7–14 are the map of what production-grade looks like — the gates, the pipeline, the judgment calls. They explain what to build and why, and point to the tools; they are not step-by-step installers for a full deployment stack.