Skip to content

AI Developer Tools · Free preview

Claude Code and Codex: Agentic Coding in Practice

Direct Claude Code and Codex on a real TypeScript app with 14 backlog items and acceptance tests. Plan multi-file changes, fix bugs test-first, prove UI fixes in a browser and review every diff; then add AGENTS.md, skills, subagents, hooks, MCP servers and plugins, and scale out with worktrees, cloud agents, CI and the SDKs.

Intermediate · professional developers
About 275 min reading
28 authored sections

Before you start: Comfortable with Git, a terminal and JavaScript or TypeScript, and Node.js 22.18 or newer. The practice app runs locally with no paid services; using the agents needs your own Claude or ChatGPT plan, or API keys.

Plan for 20–30 hours of hands-on work in addition to reading. This is an initial estimate; setup and extensions vary.

Open before purchase

A useful preview, not a teaser

Read these complete sections to judge the technical depth, teaching style, and architecture standard before buying.

Preview 1

What you will be able to do

Free preview · you need Git, a terminal and some JavaScript or TypeScript

Preview 2

Install the tools and choose where to run them

Free preview · terminal, IDE, desktop and cloud all run the same agent

Preview 3

Build a safety net before the first edit

Free preview · Git, permission modes, the sandbox and rewind

Start the free preview

Complete curriculum

28 sections · 14 backlog items with acceptance tests + capstone

Part 1 · Get started

  1. What you will be able to doFree
  2. Install the tools and choose where to run themFree
  3. Build a safety net before the first editFree

Part 2 · The core loop on TeamBoard

  1. Map TeamBoard before you change itCourse Pass
  2. Plan a multi-file feature before any code changesCourse Pass
  3. Build the feature from the plan, and steer while it worksCourse Pass
  4. Fix a bug test-first, so the test cannot share the bugCourse Pass
  5. Prove a UI fix in a real browserCourse Pass
  6. Review the diff, then commit and open a pull requestCourse Pass
  7. Debug from measurements, not from guessesCourse Pass

Part 3 · Context, models and cost

  1. Write the instructions every session starts withCourse Pass
  2. Keep the context window working for youCourse Pass
  3. Choose the model, effort and plan for the taskCourse Pass

Part 4 · Extend the agents

  1. Turn a workflow you repeat into a skillCourse Pass
  2. Delegate to subagents with narrow jobsCourse Pass
  3. Enforce rules with hooksCourse Pass
  4. Connect outside tools through MCPCourse Pass
  5. Package your setup as a plugin your team installs in one commandCourse Pass
  6. Set permissions and the sandbox for everyday workCourse Pass

Part 5 · Scale out

  1. Run agents in parallel without collisionsCourse Pass
  2. Hand long tasks to cloud agentsCourse Pass
  3. Run agents without a person at the keyboardCourse Pass
  4. Build your own agent with the SDKsCourse Pass

Part 6 · Judgment and capstone

  1. Compare the tools on your own workCourse Pass
  2. Recognize how agent work goes wrongCourse Pass
  3. Capstone: ship saved filters from spec to pull requestCourse Pass
  4. Check what you can now doCourse Pass
  5. References and how to keep this currentCourse Pass