LLearnAI
Courses / Supplementary courses
⌨️Advanced

Claude Code

Put an agent in your terminal that reads, edits, runs and ships code with you.

Claude Code is Anthropic's agentic coding tool. It works in your terminal and editor, reads your repository, edits files, runs commands and drives git. This course installs it, teaches you to brief and review it well, extends it with MCP servers, skills, subagents and hooks, and finishes with headless use, safety controls and a capstone that automates a real repository chore end to end.

Recommended before this course: AI Power User

What you will be able to do

  • Install Claude Code, sign in, and run a productive first session in an existing repository
  • Write a CLAUDE.md and a task brief that make Claude Code work the way your project needs
  • Use plan mode, permission modes and diff review to stay in control of every change
  • Extend it with MCP servers, skills, subagents and hooks configured in settings.json
  • Run it headless in scripts and CI, and know when the Claude Agent SDK is the better tool
  • Apply safety, context and cost controls so long sessions stay safe, focused and affordable

Course outline

1

Install and first session

4 lessons · quiz

What Claude Code is, how to install and sign in, and how a first session feels.

  • What Claude Code is 10 min
  • Installing and signing in 10 min
  • Your first session: ask, read, edit, run 15 min
  • Exercise: your first three tasks exercise · 25 min
  • Module quiz 6 questions
2

Working well with it

5 lessons · quiz

Teach it your project, choose the right permission mode, brief it properly and review like a lead engineer.

  • CLAUDE.md and /init: teaching it your project 12 min
  • Plan mode, permission modes and auto-accept 12 min
  • Briefing well: context, constraints and done criteria 12 min
  • Reviewing diffs, tests and the git workflow 12 min
  • Exercise: plan, then build, then review exercise · 30 min
  • Module quiz 6 questions
3

Extending Claude Code

5 lessons · quiz

Connect it to external systems with MCP, package know-how as skills, delegate to subagents, and enforce rules with hooks.

  • MCP servers: claude mcp add 12 min
  • Skills and slash commands 12 min
  • Subagents and parallel work 12 min
  • Hooks and settings.json 14 min
  • Exercise: add one MCP server and one skill exercise · 30 min
  • Module quiz 6 questions
4

Advanced patterns

5 lessons · quiz

Keep long sessions sharp, run it without a human at the keyboard, stay safe, and manage cost and model choice.

  • Context management: /compact, /clear and long sessions 12 min
  • Headless and SDK use 14 min
  • Safety: sandboxes, secrets and destructive commands 14 min
  • Cost and model choice 12 min
  • Capstone: automate a repo chore end to end exercise · 60 min
  • Module quiz 6 questions