Topolearn

Learn Claude Code

Anthropic's agentic coding tool for the terminal, IDE, desktop and web

27 lessons · 6 hours · 4 levels · covers Claude Code 2.1.x (as of Oct 2026)

Claude Code is Anthropic's AI coding agent. It reads your codebase, edits files, runs commands and works with git, from a terminal CLI, IDE extensions, a desktop app or the web, and it can be scripted in CI or embedded through the Claude Agent SDK. Learning it well means learning to steer an agent: context, permissions, verification and extension points such as skills, hooks and MCP.

Sign in free to track progressOfficial docs ↗

Before you start

Claude Code builds on these; the tree links to the exact lessons you need.

BashEssentials levelAI & LLM fundamentalsPractical levelGitEssentials level

How deep do you want to go?

1. EssentialsRead and write simple code, do the everyday basics8 lessons · 1 hour in total
2. PracticalUse it productively at work, with the usual idioms17 lessons · 3 hours in total
3. AdvancedThe less common features, performance and trade-offs23 lessons · 5 hours in total
4. ExpertInternals and the edge cases experts know27 lessons · 6 hours in total

The Claude Code tree

Claude CodeWhat Claude Code isWhat Claude Code isInstalling and signing inInstalling and signinginYour first sessionYour first sessionHow the agent loop worksHow the agent loopworksPermission modesPermission modesSlash commandsSlash commandsCLAUDE.md memory filesCLAUDE.md memoryfilesCommits and pull requestsCommits and pullrequestsPlan modePlan modeManaging context and sessionsManaging contextand sessionsPermission rulesPermission rulesSettings files and scopesSettings files andscopesVS Code and JetBrainsVS Code andJetBrainsCost and usage awarenessCost and usageawarenessSkills and custom commandsSkills and customcommandsConnecting MCP serversConnecting MCPserversExplore, plan, code, verifyExplore, plan, code,verifyThe shell and running commandsThe shell and runningcommandsTool calling (function calling)Tool calling (functioncalling)What an AI agent isWhat an AI agent isStaging and committingStaging andcommittingThe context windowThe context windowTokensTokens
Essentials and Practical shown (17 of 27 lessons); the full list is below. Click a lesson to read it.

All 27 lessons, in order

Getting started

  1. What Claude Code is · Essentials · 6 min

    Claude Code is Anthropic's coding agent: it reads your project, runs commands and edits files while you watch and steer.

  2. Installing and signing in · Essentials · 8 min

    Install the claude command with the native installer, Homebrew or WinGet, start it in a project and log in.

  3. Your first session · Essentials · 12 min

    Ask questions about your code, point at files with @, approve or reject edits, stop Claude with Esc and rewind with Esc Esc.

  4. VS Code and JetBrains · Practical · 10 min

    Use Claude Code inside your editor with the VS Code extension or JetBrains plugin, or connect the terminal CLI with /ide.

Working with the agent

  1. How the agent loop works · Essentials · 10 min

    Claude gathers context, takes action and verifies results in a loop, choosing tools step by step until the task is done.

  2. Permission modes · Essentials · 12 min

    Choose how much Claude may do without asking: Manual, Accept edits, Plan and Auto, switched with Shift+Tab.

  3. Slash commands · Essentials · 8 min

    Built-in commands such as /help, /init, /model, /clear, /compact, /context, /usage and /rewind control the session itself.

  4. Commits and pull requests · Essentials · 10 min

    Let Claude stage, write commit messages, create branches and open pull requests, while you keep control of what is pushed.

  5. Plan mode · Practical · 10 min

    Have Claude research and write a plan you approve before it edits anything, entered with Shift+Tab or /plan.

  6. Managing context and sessions · Practical · 12 min

    Keep the context window lean with /context, /clear, /compact and /btw, and resume named sessions instead of one endless chat.

  7. Cost and usage awareness · Practical · 10 min

    Watch usage with /usage and /context, and cut tokens by clearing context, choosing models and effort, and writing specific prompts.

Memory, permissions and settings

  1. CLAUDE.md memory files · Essentials · 12 min

    Give Claude persistent project instructions in CLAUDE.md, created with /init and edited with /memory.

  2. Permission rules · Practical · 14 min

    Pre-approve, force a prompt for, or block specific tools and commands with allow, ask and deny rules.

  3. Settings files and scopes · Practical · 10 min

    Know the user, project, local and managed settings files, which one wins, and what to commit to git.

  4. Rolling Claude Code out to a team · Expert · 14 min

    Share one reviewed setup through the repo's .claude/ folder, scope rules to paths, and use managed settings for policies nobody can override.

Extending Claude Code

  1. Skills and custom commands · Practical · 14 min

    Package repeatable workflows and team knowledge as skills in .claude/skills/<name>/SKILL.md, run with /name or loaded automatically.

  2. Connecting MCP servers · Practical · 14 min

    Give Claude new tools such as issue trackers, databases or design files by adding MCP servers with claude mcp add.

  3. Subagents · Advanced · 14 min

    Delegate focused work to subagents that run in their own context window with their own tools and model, returning only a summary.

  4. Hooks · Advanced · 16 min

    Run your own shell commands at fixed points such as before a tool runs or after a file edit, to format, check or block actions.

  5. Plugins and marketplaces · Advanced · 10 min

    Install bundles of skills, subagents, hooks and MCP servers from marketplaces with /plugin, and choose user, project or local scope.

Effective workflows

  1. Explore, plan, code, verify · Practical · 14 min

    The core workflow: explore read-only, plan, implement against a check such as tests, then commit, with specific prompts throughout.

  2. Parallel sessions and scaling up · Expert · 16 min

    Run several Claude sessions in git worktrees, use writer/reviewer pairs, fan out claude -p over many files, and gate unattended runs with checks.

Security and isolation

  1. Security and prompt injection · Advanced · 16 min

    Understand prompt injection and the risks of an agent with shell access, and know what never to auto-approve.

  2. Sandboxing and auto mode · Expert · 16 min

    Enforce OS-level filesystem and network limits on shell commands with /sandbox, and know what auto mode and bypassPermissions do and don't protect.

Automation, CI and SDK

  1. Headless mode and scripting · Advanced · 14 min

    Run Claude non-interactively with claude -p, pipe data in, get text or JSON out, and pre-approve exactly the tools a script needs.

  2. Claude in GitHub Actions · Advanced · 14 min

    Run Claude Code in GitHub workflows with anthropics/claude-code-action to answer @claude mentions or run prompts on events.

  3. The Claude Agent SDK · Expert · 16 min

    Embed Claude Code's agent loop, tools, permissions and hooks in your own Python or TypeScript program with the Agent SDK.