Claude Code Mastery: From Beginner to Expert
Stanshy | 2026-03
This course is built on David Chu's Five-Layer Evolution Model, systematically guiding readers from Claude Code fundamentals all the way to Harness Engineering and autonomous Agent systems.
All chapters revolve around a single running case project (AI Dev Assistant), ensuring every concept has actionable, hands-on examples.
Module
Evolution Layer
Topic
Chapters
1
Layer 1: Conversational
Introduction to Claude Code
2
2
Layer 2: Configured
Configuration
3
3
Layer 3: Automated
Hooks / Skills / MCP
4
4
Layer 4: Orchestrated
Subagents / Context / Parallel Dev
3
5
Layer 5: Autonomous
Agent Teams / Headless / SDK
3
6
Methodology
Harness Engineering
2
7
Case Studies
Real-World Frameworks
4
8
Reference
Cheatsheet, Resource Index, Disaster Recovery
3
Module 1: Introduction to Claude Code
Chapter
Title
Summary
01-1 What is Claude Code
Agentic CLI definition, Agent Loop, tool comparison
01-2 Installation & First Success
Three installation methods, authentication, project scaffolding
Chapter
Title
Summary
02-1 CLAUDE.md Complete Guide
Boris Cherny philosophy, writing standards, size control
02-2 Settings & Permissions
settings.json layers, five permission modes, allow/deny syntax
02-3 Agent = Model + Harness
Harness core concept preview, with vs without Harness comparison
Chapter
Title
Summary
03-1 Hooks Basics
Deterministic vs advisory, 6 core events, configuration format
03-2 Advanced Hooks & Validator Pattern
Stop Hook Validator Pattern, four handler types
03-3 Skills System
On-demand loading, SKILL.md format, dynamic content injection
03-4 MCP Servers
External tool protocol, GitHub MCP, Tool Search
Chapter
Title
Summary
04-1 Custom Subagents
Independent context, role definition, three invocation methods
04-2 Context Management Strategies
8 management techniques, Document & Clear method
04-3 Parallel Development & Worktree
Git Worktree, Boris's 5+5 parallel mode
Chapter
Title
Summary
05-1 Agent Teams
Team Lead + Teammates collaboration architecture
05-2 Headless Mode & CI/CD Integration
Non-interactive mode, GitHub Actions automation
05-3 Agent SDK
Python/TypeScript SDK programmatic control
Module 6: Harness Engineering
Chapter
Title
Summary
06-1 Seven Design Principles
Context architecture, constraints, self-validation, isolation, entropy, modularity, agent readability
06-2 Build a Harness from Scratch
Pitfall-driven six-step process, three anti-patterns, full evolution example
Module 7: Real-World Frameworks
Chapter
Title
Summary
07-1 Boris Cherny Workflow
259 PRs in 30 days, 5+5 parallel mode, CLAUDE.md institutional memory
07-2 gstack Role System
YC president's 18-skill engineering team simulation
07-3 Open-Source Harness Cases
Three top project design analyses, seven-concept cross-reference
07-4 David Chu Evolution Framework
Five-layer model, three graduation paths, Boris vs RPI integration
Chapter
Title
Summary
08-1 Cheatsheet
Commands, shortcuts, Hook events, permission syntax, troubleshooting
08-2 Resource Index
Curated 37 resources, Top 10 must-bookmarks
08-3 Disaster Recovery & Debug Guide
5 fault modes, debugging workflows, --resume complete usage
Recommended Reading Order
Background
Suggested Path
Never used Claude Code
Module 1 → 2 → 3 → 4 → 5
Know basics, want to level up
Module 3 → 4 → 6 → 7
Want to learn Harness Engineering
Module 6 → 7 → 3
Limited time, just 3 chapters
01-1 → 02-1 → 03-2