For beginners who can do basic coding and command line work but have not used Claude Code systematically yet, this course helps you understand what Claude Code is, how to prepare your environment, how to start a first session, how to let it understand the project first, and how to complete one safe code change with permission control, validation, and human review.
适合人群
Developers who know a little code editing and basic command line, and want to try Claude Code for the first time
People who have used Claude or ChatGPT for code before, but are not yet comfortable with AI coding agents
Learners who want AI to read projects, locate code, make small changes, and run validation
Developers who worry that Claude Code might change files unexpectedly or bypass risk boundaries
解决的问题
Not sure what differentiates Claude Code from chat-based coding assistants
Unclear which account, terminal, editor, practice project, and git state are needed to start
Tendency to ask Claude Code to modify code immediately in the first session without a clear scope
Unclear how to ask Claude Code to first read the project, provide rationale, and propose a reviewable plan
Unsure how to inspect diff, run validation, recover from failure, and control high-risk actions
学完收获
Can explain, in your own words, what Claude Code is and how it differs from ordinary chat-based AI coding help
Can check their onboarding environment, choose a low-risk practice project, and ensure a reversible state
Can launch Claude Code in a project directory and complete a first session that only reads the project
Can explain how context, tools, and permissions together define what Claude Code can and cannot do
Can ask Claude Code to output a project understanding note and a reviewable change plan
Can complete one small code change with diff review, validation, human review, and clear outcome summary
课程亮点
Targets Claude Code beginners and avoids presenting it as a general chat-based coding question tool
Built around real project workflows: project reading, planning, incremental changes, and verification
Emphasizes permissions, Git, checkpoints, tests, and human review
Shows clear decision boundaries around CLAUDE.md, hooks, MCP, and subagents
Each lesson includes explicit judgment checkpoints for AI-led follow-up and diagnostics
全部章节
9 个章节
01
Lesson 1: Meet Claude Code as an AI Coding Agent
Explain what Claude Code is in your own words and distinguish it from ordinary chat-based AI coding assistants.
02
Lesson 2: Prepare Your Account, Terminal, Editor, and Practice Project
Determine whether you have the minimum onboarding setup for Claude Code and explain the role of each preparation item.
03
Lesson 3: Start Claude Code and Let It Understand the Project First
Start Claude Code in the correct project directory and complete a first session that only understands the project without editing any files.
04
Lesson 4: Understand Context, Tools, and Permissions
Explain how context, tools, and permissions together determine what Claude Code can see, what it can do, and when review is required.
05
Lesson 5: Read the Project First
Have Claude Code build a project map and output notes covering project overview, structure map, core flow, verification approach, and beginner task candidates.
06
Lesson 6: Ask Claude Code to Plan Before Editing
Output a reviewable plan from Claude Code before code changes and evaluate whether the plan is clear, safe, and verifiable.
07
Lesson 7: Make One Small Change and Verify It
Use an approved plan to authorize Claude Code for one small change, then judge acceptance by reading diff, running validation, and summarizing.
08
Lesson 8: Control Risk with Permissions, Testing, and Human Review
Choose permissions, validation, and review mode based on the risk level of Claude Code actions, and avoid auto-approving high-risk actions as a beginner.
09
Lesson 9: Choose Your Next Advanced Path with Claude Code
Review the beginner collaboration flow and choose an appropriate advanced path based on real development needs.
