Claude Code 101
A hands-on introduction to Claude Code — Anthropic's agentic coding CLI — covering its core tool loop, safe permission model, and how to direct it effectively on real codebases.
Overview
This foundational course walks through how Claude Code actually works under the hood: the agentic loop of reading files, planning a change, editing code, running shell commands, and verifying the result before reporting back. It covers the core tool set (Read, Edit, Write, Bash, Grep, Glob), the permission model that gates risky actions behind confirmation, and how to scope a prompt so the agent solves the right problem instead of guessing at one.
What I Learned
- How Claude Code's agentic loop works: reading files, planning edits, running shell commands, and verifying results before reporting back
- The core toolset (Read, Edit, Write, Bash, Grep, Glob) and when to reach for each one instead of a generic shell command
- How the permission system balances autonomy with safety — auto-approved reads vs. confirmation-gated writes and destructive commands
- Writing a CLAUDE.md / AGENTS.md file to give a project persistent context: conventions, docs to check first, anything that breaks default assumptions
- Breaking a vague request into a concrete, verifiable plan before any code gets touched
- Reviewing the diffs and commit messages Claude Code produces the way I'd review a teammate's PR, not rubber-stamping them
How I'm Using It in Real Development
This portfolio's own AGENTS.md file is the first thing Claude Code reads before touching the repo — it flags that this Next.js setup has framework-level differences from training data and points to the bundled docs to check first
Day-to-day I use it for the small, time-consuming stuff: fixing a scroll-jank bug, wiring up a new npm package, cleaning up an unused import — places where a fast, correct first pass matters more than deep architecture
I use the permission prompts deliberately rather than blanket-approving everything: file edits and local build/lint commands run freely, while git pushes and destructive resets stay gated behind an explicit yes
- Issuer
- Anthropic
- Issued
- April 24, 2026
- Certificate No.
- wmhrfygd6fiq