Skip to content
All courses

06 · Rules · small loops · handoffs

Build Apps with AI Coding Agents

Work with coding agents like Claude Code, Cursor or Codex the way a careful team lead would: project rules, small tasks with a clear finish line, reviewed diffs, local tests before every push, and a handoff file so no context is lost.

Beginner to intermediate · 3 modules · 7 lessons · about 39 min

Modules

  1. 01Set up2 lessons
  2. 02Work in small loops2 lessons
  3. 03Keep it safe3 lessons

What you’ll be able to do

  • 1Write an AGENTS.md that keeps every agent on your project's rules
  • 2Give tasks with a goal, scope and a clear 'done when'
  • 3Review an agent's diff before it reaches git
  • 4Run checks locally so CI minutes aren't wasted
  • 5Keep a HANDOFF.md so any agent or person can continue the work
  1. Module 01

    Set up

    What agents do, and the rules file that steers them.

    1. 01How coding agents workThey read files, run commands and edit code. You stay the reviewer.5 min
    2. 02Project rules: AGENTS.mdOne file every agent reads first, so you never repeat yourself.7 min
  2. Module 02

    Work in small loops

    Clear tasks in, reviewed diffs out.

    1. 03Small tasks with a clear 'done'A good task names the goal, the files and how you'll know it worked.6 min
    2. 04Review every diffFive minutes of reading saves a week of debugging.5 min
  3. Module 03

    Keep it safe

    Local checks, handoffs and secrets.

    1. 05Test before you pushLocal checks are free; cloud minutes are not.5 min
    2. 06HANDOFF.md: never lose contextA living file that lets any agent or person pick up the work.6 min
    3. 07Secrets and permissionsWhat an agent should never see or do.5 min