

Outline:
Building the right AI-coding workflow decides whether you ship clean, production-grade code or drown in a tangled 'spaghetti' codebase even you can't read. This course teaches you to use Claude Code as a professional tool, not a crutch you lean on.
It covers how large language models work, why context windows matter, and how to manage them without losing control of the codebase. All lessons are released asynchronously every Monday; live Q&A spans multiple time zones. You keep all the learning materials after the course.
What you'll learn:
- How to use the plan / execute / clean loop to keep Claude Code in top shape
- How to write AGENTS.md files and custom skills to steer output without wasting tokens
- How to plan and execute features bigger than a single model session via multi-stage plans and the 'tracer bullet' method
- How to build automated feedback loops — tests, pre-commit hooks and red-green refactoring — catching issues before they pile up
- How to use the Ralph loop to let Claude Code work autonomously through a GitHub Issues todo list
Tutorial language: bilingual subtitles / Chinese-English
Docs language: English / translate with AI yourself
Security scan: no virus no plugin / cloud scan Virustotal Virscan
File format: video / docs / image-text
Compression: 7ZIP
Video player: Perfect Codec / potplayer / QQ Player
Course catalog (note: source titles are original English; docs not yet translated, use AI to translate):
│ code.zip
│ lesson1
│ lesson5
│ lesson6
│ lesson7
│ lesson8
│ lesson9
│ lesson10
│ lesson11
│ lesson12
│ lesson13
│ lesson14
│ lesson15
│ lesson16
│ lesson17
│ lesson18
│ lesson19
│ lesson20
│ lesson21
│ lesson22
│ lesson23
│ lesson24
│ lesson25
│ lesson26
│ lesson27
│ lesson28
│ lesson29
│ lesson31
│ lesson32
│ lesson33
│ lesson34
│ lesson35
│ lesson36
│ lesson37
│ lesson38
│ lesson39
│ lesson40
│ lesson41
│ lesson42
│ lesson43
│ lesson44
│ lesson45
This course comes with a materials pack. Members receive the access link by email after requesting the course.