Course catalog
Set the rules, review the work, and run several coding agents in parallel.
Coding With AI Agents taught you the loop: brief, plan, apply, verify. This course is the day job. You'll learn the tool that runs the loop for you — the harness — and the five families they come in, from terminal agents like Claude Code, Codex CLI and opencode to review bots and background runs. You'll read a session the way a pilot reads an instrument panel, choose permission modes and sandboxes on purpose, and engineer the context an agent gets instead of dumping your repo into it. Then the part that separates a fast week from a wasted one: get the plan before the diff, triage a 400-line change in three passes, make a failing test the contract, and calibrate how hard to verify by how hard it is to undo. You'll finish running more than one agent at a time — subagents, git worktrees, and hooks that block a bad command instead of hoping nobody typed it.
Section 1
The program that runs the agent loop for you — how to read it, and how to set its rules.
Section 2
The agent only knows what it can see. Decide what it sees, and when to start again.
Build and run a real AI product: it searches your own content, uses only the tools you approved, is tested the same way every time, has a plan for when it fails, and shows its speed and cost.
11 course sequence
Ship a secure TypeScript app built with AI in your workflow: tested algorithms, tools kept inside safe limits, architecture choices you can explain, and releases you can check after they go out.
16 course sequence
7-day free trial on every plan for first-time members · Cancel anytime
Get startedSection 3
Get the plan before the diff, triage the diff fast, and let a test carry the argument.
Section 4
More than one agent, more than one checkout, and the automation that keeps them honest.