Learn Claude Code by doing, not reading
TL;DR Highlight
An interactive Claude Code learning platform featuring a browser-based terminal simulator, Config Builder, quizzes, and more — letting you practice core Claude Code features without any installation or API key.
Who Should Read
Developers new to Claude Code or those looking to systematically learn advanced features like slash commands, hooks, CLAUDE.md, and MCP server configuration. Especially suitable for beginners who want to explore the features before installing anything.
Core Mechanics
- claude.nagdy.me, created by Ahmed Nagdy, is an interactive learning platform for Claude Code that runs entirely in the browser — no API key or separate installation required.
- The platform consists of 11 modules ranging from Beginner to Advanced, progressing through Slash Commands (30 min), Memory & CLAUDE.md (45 min), Project Setup (45 min), Commands Deep Dive (30 min), Skills (1 hour), and more.
- A built-in browser terminal simulator lets you practice features like slash commands, hooks, and skills by actually typing them out, giving you a feel for the real environment before you ever install anything.
- The Config Builder tool lets you interactively generate CLAUDE.md files (per-project AI behavior instruction files), hooks (scripts that respond to specific events), MCP server configurations, and plugin settings via a form — then copy and paste them directly into your project.
- Each module ends with a quiz that doesn't just reveal the correct answer when you're wrong — it explains why that answer is correct, helping you solidify your understanding of the concepts.
- Additional reference tools beyond the core curriculum are also provided, including a Playground (dedicated terminal sandbox), a Cheat Sheet (summary of frequently used commands and shortcuts), and a Feature Index (searchable list of all features).
- A 'Find Your Level' quiz is designed to diagnose your skill level and direct you to the appropriate starting module. However, as community feedback suggests, the accuracy of this assessment has been called into question.
Evidence
- "Several advanced users who use Claude Code daily — including plugins, skills, MCP servers, and subagent workflows — reported being rated as Beginners after taking the 'Find Your Level' quiz. There are reports that selecting D or C, or even all D answers, still results in a Beginner rating, suggesting a bug or design flaw in the quiz logic. Some developers familiar with terminals argued that just installing Claude Code and using it directly is faster and more genuine 'learning by doing,' and while the platform may be more useful for complete beginners who find terminals intimidating, those users would actually need even more foundational explanations of what a terminal and commands are. Comments also raised concerns about how quickly Claude Code burns through quota — one user on the Max5 plan ($100/month) reportedly consumed roughly 10% of their session quota in just 10 minutes with a single prompt, with suspicion that Anthropic recently changed something. Separate from the learning platform itself, this raises practical concerns about cost predictability for those considering adopting Claude Code. Some comments expressed fatigue at having to learn yet another tool, questioning whether it's worth learning Claude Code syntax when you can just tell the AI what you want in plain language — though the implicit counterargument is that knowing how to use structured features like slash commands and hooks precisely does produce better results in practice. There was also a cynical 'RTFM' reaction criticizing people for seeking a learning platform instead of reading the official docs, while others countered that interactive learning is genuinely faster than reading documentation — highlighting a difference in learning style preferences."
How to Apply
- "If your team is adopting Claude Code for the first time, try using the Config Builder at claude.nagdy.me to interactively generate a draft CLAUDE.md file. Enter your project information into the form and get a ready-to-copy config file instantly — a much faster starting point than writing one from scratch. If you want to try hooks or MCP servers in Claude Code but find the setup daunting, use the platform's terminal simulator to familiarize yourself with the command flow before installing anything — this can help reduce mistakes in the real environment. Hooks configuration in particular can break your entire workflow if misconfigured, so prior practice is valuable. For teams that need to onboard members to Claude Code, the platform's 11-module sequence (Slash Commands → Memory & CLAUDE.md → Project Setup → Commands Deep Dive → Skills) can be used directly as a team learning curriculum. Each module includes an estimated completion time, making it easy to plan a schedule. If you're evaluating Claude Code on the Max5 plan at $100/month, be aware that community reports suggest quota can be consumed much faster than expected — it's worth monitoring quota usage by usage pattern early on and building a cost prediction model."
Terminology
CLAUDE.mdA configuration file that Claude Code reads from the project directory, serving as a kind of 'README for the AI' — pre-specifying how the AI should behave in that project, including coding style, prohibited actions, and project structure.
slash commandsShortcut commands in Claude Code that begin with /. For example, /help or /clear — used to quickly invoke specific functions.
hooksScripts that run automatically when a specific event occurs. For example, automatically running a linter after a file is saved, or inserting specific processing before Claude responds.
MCPShort for Model Context Protocol — a standard protocol created by Anthropic that allows AI models to communicate with external tools (file systems, databases, APIs, etc.). Configuring an MCP server enables Claude Code to use those tools directly.
skillsReusable procedures in Claude Code that pre-define frequently repeated tasks or specific patterns — essentially custom macros or workflow templates.
subagent workflowAn approach in Claude Code where a single large task is split across multiple sub-AI agents for processing. Complex tasks can be handled in parallel or sequentially, enabling more efficient execution.