Claude Code SELF
This is a course about Claude Code — a tool that works with any stack and any task. You study the agent itself in a comic-story format, and every task is a training ground for the real workflow. What matters here are the skills of working with an AI agent: writing specs, reviewing diffs, debugging with evidence, keeping a safe git flow, and managing context. And supplementary materials on development basics give you the foundation to understand how it all works. Go from a simple prompt to the discipline of "plan - edit - verify - ship" and become an AI-native developer.
- Level 1
Unavailable Claude Code: an agent, not a chat - Level 2
Unavailable Claude Code in the IDE - Level 3
Unavailable Slash commands and session rewind - Level 4
Unavailable Skills: ready-made agent capabilities - Level 5
Unavailable Plugins: bundles of capabilities - Level 6
Unavailable Subagents: a second worker - Level 7
Unavailable MCP: a bridge to live data - Level 8
Unavailable Hooks: event-driven automation - Level 9
Unavailable Autonomous agent: Playwright and /goal - Level 10
Unavailable CLAUDE.md: project context and memory - Level 11
Unavailable The cost of agent work: /model, tokens and /context - Level 12
Unavailable Plan mode and permissions - Level 13
Unavailable Change control: git and /diff - Level 14
Unavailable Prove you fixed it: logs, tests and /model - Level 15
Unavailable Reviewing changes: /diff and scope - Level 16
Unavailable Context: /compact and plan mode