Claude Code Cheat Sheet, Everything-Claude-Code & Claude How-To — Complete Reference Kit

Claude Code Cheat Sheet, Everything-Claude-Code & Claude How-To — Complete Reference Kit

What Are These Resources?

Three complementary references that together form a complete Claude Code toolkit:

  1. Claude Code Cheat Sheet (v2.1.61) — A dense, visual one-page reference covering every major feature of Claude Code: keyboard shortcuts, slash commands, MCP servers, memory files, workflows, configuration, and agent patterns. Designed to be printed or pinned as a quick-lookup companion.

  2. everything-claude-code — An open-source GitHub repository (github.com/affaan-m/everything-claude-code) that exploded to 120K+ stars within days of release. It provides a complete, battle-tested Claude Code configuration — from agent setup to skill definitions, shortcuts, rules, and MCP connections. The configs won a hackathon and work on both Windows and Mac.

  3. Claude How-To — A visual, example-driven guide (github.com/luongnv89/claude-howto) with 13K+ stars that takes you from zero to advanced in a weekend. 10 progressive modules, copy-paste templates, and a structured learning path. v2.2.0, MIT licensed.

Together: the cheat sheet tells you what exists, the everything repo shows you how to wire it all up, and Claude How-To teaches you step by step.

*Source: Weibo post by 默庵·超级个体 Claude How-To*

Cheat Sheet — Key Sections at a Glance

Keyboard Shortcuts

Shortcut Action
Esc Cancel current generation
Ctrl+C Interrupt / clear input
Ctrl+L Clear conversation
Ctrl+R Fuzzy search conversation history
Tab Accept autocomplete suggestion
Shift+Tab Toggle between compact/full mode
Ctrl+J Newline in input (multi-line editing)
Ctrl+K Toggle context panel

Slash Commands

Command Purpose
/help Show available commands and usage
/clear Clear conversation context
/compact Summarize and compact the conversation
/review Trigger code review agent
/init Generate a CLAUDE.md for the current project
/cost Show token usage and cost for the session
/config Open or edit configuration
/memory View and manage memory files
/permissions Review and adjust tool permissions
/vim Toggle Vim keybindings

MCP Servers

The cheat sheet covers how Claude Code connects to Model Context Protocol (MCP) servers — external tool providers that extend Claude’s capabilities (browser automation, database access, GitHub, custom APIs). Key points:

  • MCP servers are configured in .claude/settings.json or project-level claude.json
  • Each server exposes tools that Claude can call during a session
  • Common MCP servers: Playwright (browser), GitHub, filesystem, database connectors
  • Servers can be local processes or remote endpoints

Memory & Files

Claude Code uses a layered memory system:

File Scope Purpose
~/.claude/CLAUDE.md Global (all projects) Personal preferences, workflow rules, global instructions
./CLAUDE.md Project root Project-specific context, conventions, architecture notes
.claude/settings.json Project Tool permissions, MCP server config
memory/ directory Project Persistent memory files that survive across sessions — auto-updated by Claude

The cheat sheet emphasizes: CLAUDE.md is your highest-leverage file. It overrides default behavior and shapes every interaction.

Workflows & Tips

  • Plan Mode — Ask Claude to plan before executing. Use “think hard” or “plan first” to trigger deeper reasoning without immediate code changes.
  • Permissions — Claude asks before file writes and shell commands by default. Use /permissions or allowedTools in settings to pre-approve trusted operations.
  • Voice Mode — Dictate tasks instead of typing. Useful for brainstorming or when away from keyboard.
  • Sparse Checkouts — For monorepos, use Git sparse-checkout so Claude only sees relevant subdirectories, reducing context and cost.
  • Headless Mode — Run Claude Code non-interactively in CI/CD pipelines with claude -p "prompt".

Config & Environment Variables

Variable / Flag Purpose
ANTHROPIC_API_KEY API key for direct API access
CLAUDE_CODE_USE_BEDROCK=1 Route through AWS Bedrock
CLAUDE_CODE_USE_VERTEX=1 Route through Google Vertex
--model Override default model
--max-turns Limit agentic loop iterations
--allowedTools Pre-approve specific tools
--output-format Set output format (json, text, stream-json)

Agents & Advanced Patterns

The cheat sheet covers several agent patterns:

  • Parallel Code Review — Spawn sub-agents that review different files simultaneously, then merge findings.
  • Simplify Pattern — Use --allowedTools to restrict an agent to read-only tools for analysis tasks.
  • Skill Paths — Define reusable skill files (.md instructions) that Claude loads on demand via slash commands.
  • Orchestrator Pattern — A main agent delegates subtasks to specialized sub-agents, each with their own tool permissions and context.

Everything-Claude-Code Repository

Repo: github.com/affaan-m/everything-claude-code

What It Provides

A complete, ready-to-use Claude Code configuration that covers the full setup journey:

Component What You Get
Agent configs Pre-built agent definitions for code review, refactoring, documentation, testing
Skills library Reusable .md skill files for common workflows (commit, PR review, deploy, etc.)
Shortcut mappings Custom slash commands mapped to complex multi-step workflows
Rules & conventions CLAUDE.md templates with battle-tested instruction patterns
MCP connections Ready-made MCP server configurations for popular tools
Cross-platform Tested configs for both Windows and macOS

Why It Went Viral

  • Saves hours of manual configuration — copy, adapt, run
  • Real-world tested (hackathon-winning setup)
  • Covers the full spectrum from beginner (“what’s a CLAUDE.md?”) to advanced (parallel agent orchestration)
  • Community-driven with rapid iteration

Repository Structure (Key Directories)

everything-claude-code/
├── agents/          # Agent definitions and orchestration configs
├── skills/          # Reusable skill files (.md)
├── rules/           # CLAUDE.md templates and project rules
├── mcp/             # MCP server configurations
├── shortcuts/       # Custom command definitions
├── examples/        # Worked examples and use cases
└── docs/            # Setup guides for Windows and Mac

Claude How-To — Visual Learning Path

Repo: github.com/luongnv89/claude-howto (13.1K stars, 1.4K forks, v2.2.0)

The Problem It Solves

You installed Claude Code, ran a few prompts — now what? The official docs describe features but don’t show how to combine them. There’s no clear learning path. Examples are too basic to build production pipelines. Claude How-To fills that gap with visual tutorials and copy-paste templates.

10 Progressive Modules

Module Focus Level Duration
Slash Commands User-invoked shortcuts Beginner 30 min
Memory Persistent context Beginner+ 45 min
Checkpoints Session snapshots/rewind Intermediate 45 min
CLI Basics Terminal commands Beginner+ 30 min
Skills Reusable capabilities Intermediate 1 hour
Hooks Event-driven automation Intermediate 1 hour
MCP Protocol External tool access Intermediate+ 1 hour
Subagents Specialized AI assistants Intermediate+ 1.5 hours
Advanced Features Planning, thinking, background tasks Advanced 2-3 hours
Plugins Bundled feature collections Advanced 2 hours

Total: ~11-13 hours for the full path, but you get value in 15 minutes.

Real-World Use Cases

Use Case Features Combined
Automated Code Review Slash Commands + Subagents + Memory + MCP
Team Onboarding Memory + Slash Commands + Plugins
CI/CD Automation CLI + Hooks + Background Tasks
Documentation Generation Skills + Subagents + Plugins
Security Audits Subagents + Skills + Hooks (read-only)
Complex Refactoring Checkpoints + Planning Mode + Hooks

Feature Comparison Matrix

Feature Invocation Persistence Best Use
Slash Commands Manual (/cmd) Session only Quick shortcuts
Memory Auto-loaded Cross-session Long-term learning
Skills Auto-invoked Filesystem Automated workflows
Subagents Auto-delegated Isolated context Task distribution
MCP Protocol Auto-queried Real-time Live data access
Hooks Event-triggered Configured Automation & validation
Plugins Single command All features Complete solutions

Quick Start

git clone https://github.com/luongnv89/claude-howto.git
cd claude-howto

# Copy a slash command to your project
cp 01-slash-commands/optimize.md .claude/commands/

# Add project memory
cp 02-memory/project-CLAUDE.md ./CLAUDE.md

# Install a skill
cp -r 03-skills/code-review ~/.claude/skills/

How LearnAI Team Could Use This

  • Use the cheat sheet as a quick-reference handout for Claude Code onboarding workshops.
  • Fork the everything-claude-code repo as a starting template for team-standardized CLAUDE.md, skills, and MCP configs.
  • Assign Claude How-To modules as progressive homework for new team members learning agent workflows.
  • Build a shared team CLAUDE.md using the Boris template pattern from the everything repo.

How to Use These Resources

Quick Start (5 minutes)

  1. Bookmark the cheat sheet — Keep it open or print it. Use it as a lookup reference when you forget a shortcut or command.
  2. Clone the everything repo:
    git clone https://github.com/affaan-m/everything-claude-code.git
    
  3. Copy the CLAUDE.md template from rules/ into your project root. Edit it to match your project.
  4. Pick 2-3 skills from skills/ that match your workflow. Drop them into your project’s .claude/ directory.

Progressive Adoption

Stage Action Payoff
Week 1 Set up CLAUDE.md + learn 5 keyboard shortcuts Faster interactions, consistent behavior
Week 2 Add MCP servers (GitHub, Playwright) Claude can browse, create PRs, run tests
Week 3 Define custom skills for your repeated tasks One slash command replaces 10 minutes of prompting
Week 4 Set up parallel agents for code review Automated multi-file review on every PR

Tips for Maximum Value

  • Start with the cheat sheet to understand what’s possible, then use the everything repo to implement it.
  • Don’t copy everything at once — adopt incrementally based on your actual pain points.
  • The memory/ directory is underused by most people. Let Claude persist context between sessions and your onboarding cost drops to near zero.
  • For teams: fork the everything repo and customize it as your shared Claude Code standard.