About this course
You didn't study CS, but you want to genuinely master Claude Code. Most content teaches you to
make the AI write more code; this course does the opposite — it teaches the 20% the agent
cannot do, which decides whether you succeed.
Agents have eaten the "idea → syntactically correct code" step. What they have not eaten are
four human-machine seams: knowing what you actually want, judging whether the output is right,
the reality of your machine and environment, and the causal truth buried in a stack trace.
So this course is ordered by "what most unlocks your ability to command an agent", not by a
university syllabus. Everything happens in your own real sandbox: real terminal, real files,
real git, real errors.
What you'll learn
- Specify a task for an agent: goal, constraints, acceptance criteria, examples, and what not to touch
- Verify an agent's output: run it, read the diff, demand a test that can fail — never trust 'done'
- Make things run in a real terminal: paths, dependencies, env vars, secrets (and why secrets never enter git)
- Read errors and stack traces, find your own line, and explain the causality back to the agent
- Use git as a safety net and a review surface: commits, branches, diffs, rollback, ssh-keys
- Explain what you're building: HTTP request/response, frontend vs backend, APIs/JSON, how login works
- Treat big-O, data structures and the classic algorithms as vocabulary and taste — enough to question the agent
- Extend an agent with MCP (tools) and Skills (procedures), and know where your own moat is
Syllabus
1Before You Start: What Is AI Coding1 session
- 1What Is AI Coding? What Can It Do for You?15 minStart →
2Meta-Skills: The Highest-Leverage 20%4 sessions
- 1What Agents Have Eaten, and What They Haven't: The Four Seams15 minStart →
- 2Specify the Work: The Five Elements of a Task Spec16 minStart →
- 3Verification: It Says It's Done — On What Grounds Do You Believe It?18 minStart →
- 4DRY / KISS / YAGNI: Three Phrases for Steering and Reviewing an Agent16 minStart →
3Make It Run, Read the Feedback: You Are the Agent's Hands and Eyes8 sessions
- 1The Terminal: Where the Agent Lives16 minStart →
- 2Filesystem, Paths and the Current Directory: Half of 'I Changed It and Nothing Happened'16 minStart →
- 3The Temperament of Linux: Why Servers and Agents Live Here16 minStart →
- 4Environments, Dependencies and Package Managers: Curing 'But It Works On My Machine'17 minStart →
- 5Env Vars and Secrets: The Thing in Your Code That Must Never Enter Git17 minStart →
- 6Reading Errors, Logs and Stack Traces: What That Wall of Red Actually Says17 minStart →
- 7Ports and Processes: Is the Service Even Up?16 minStart →
- 8The Editor / IDE: Your Cockpit for Watching the Agent12 minStart →
4How to Manage Your Own Project6 sessions
- 1Git's Mental Model: Save Points and Parallel Saves15 minStart →
- 2Reading Diffs: The Main Arena for Reviewing an Agent17 minStart →
- 3.gitignore: Keep Secrets Out of Git Forever15 minStart →
- 4Undo and Rollback: One Move Back When It Goes Wrong15 minStart →
- 5PR and Code Review: The Reviewer's Posture, Even Alone14 minStart →
- 6The ssh-key: Public Key to Them, Private Key to You15 minStart →
5What Am I Building: HTTP, Frontend/Backend, APIs and Login5 sessions
- 1HTTP: One Sends a Request, One Returns a Response15 minStart →
- 2Frontend vs Backend: Where Does Your Code Actually Run15 minStart →
- 3APIs and JSON: How Programs Talk to Each Other15 minStart →
- 4Magic Link: How Login Actually Works15 minStart →
- 5The Database: Where Your Data Lives, and Who Can Read or Write It15 minStart →
6Make Programs Run Faster: Algorithms & Data Structures, Fast-Track4 sessions
- 1Big-O Intuition: Ten Times the Data, a Hundred Times Slower?14 minStart →
- 2Data-Structure Vocabulary: What list / dict / set / tree / graph Each Do14 minStart →
- 3Binary Search: Halving Your Way Through Sorted Data14 minStart →
- 4Three More Patterns, Just Know Their Faces: Graph Traversal, Divide-and-Conquer, Randomization14 minStart →
Same series · 面向非科班的快速能量补充
Fast Track for Non-CS Backgrounds
Not from a CS background? Get up to speed fast — the essentials of agents and computer science.