Claude Code for Developers: Foundations

Claude Code is an AI agent that lives in your terminal. Unlike autocomplete or a chat window, it reads your files, runs commands, writes code, and works through problems while you watch or step away. This course teaches you to use it well, from your first session to shipping a real change, one step at a time.

You will learn two things by heart:

• The Loop: the everyday rhythm of coding with an agent: Explore, Plan, Code, Verify, Commit. Skipping straight to "write the code" is the single most common beginner mistake, and the Loop is the fix.
• SPEC: how to ask for what you want so you get it: Scope the task, Point to examples, state the Expected result, and give it a Check to verify against.

What you will be able to do

• Install Claude Code, start a session, and understand the approval model that keeps you in control
• Write a CLAUDE.md so the agent knows your project
• Explore and plan a change before a single line is written
• Read a diff, approve changes, and course-correct the moment things drift
• Give the agent a way to verify its own work, so "looks done" becomes "is done"
• Commit and open a pull request, with good habits that scale

Who this course is for

Developers of any level who are new to agentic coding, plus anyone comfortable in a terminal who wants to build software with an AI agent. You do not need to be an expert. You do need a willingness to work in the command line and review what the agent does.