Claude Code Cheat Sheet

Here’s a rewritten Claude Code Cheat Sheet with relevant examples for real development scenarios, showing both command usage and prompt formats:

Claude Code Cheat Sheet (with Examples)

Quick Start

Install and Launch:

npm install -g @anthropic-ai/claude-code
claude           # Launch interactive shell
claude --version # Check your installed version        

npm install -g @anthropic-ai/claude-code claude # Launch interactive shell claude --version # Check your installed version

Use REPL for interactive sessions or CLI flags for one-off jobs.

Basic Usage Examples

Generate Feature Boilerplate

claude -p "Create a user authentication module in Python using JWT. Include signup, login, logout, and token verification functions."        

claude -p "Create a user authentication module in Python using JWT. Include signup, login, logout, and token verification functions."

Claude outputs ready-to-use Python files implementing JWT authentication.

Debug Codebase

claude -p "Diagnose and fix error: 'TypeError: unhashable type: list' found in server.py"        

claude -p "Diagnose and fix error: 'TypeError: unhashable type: list' found in server.py"

Claude finds the line, explains the cause, and suggests (or directly edits) the fix.

Planning & Architecture

Prompt:

ultrathink: Plan a scalable microservices architecture for an e-commerce platform using Node.js, MongoDB, and Docker.        

ultrathink: Plan a scalable microservices architecture for an e-commerce platform using Node.js, MongoDB, and Docker.

Claude breaks down modules, recommended data models, and container setup. Asks for feedback before proceeding.

Testing Workflow

Prompt:

Write tests for 'calculateTax' function in tax_utils.py. Use TDD: only write tests first, do not implement function yet.        

Write tests for 'calculateTax' function in tax_utils.py. Use TDD: only write tests first, do not implement function yet.

Claude provides test cases in desired framework (e.g., pytest), waits for code to pass tests before implementing.

Audit & Documentation

Prompt:

audit: Review payment_processing.js for potential security vulnerabilities (e.g., SQL injection, XSS). List each issue with a fix.        

audit: Review payment_processing.js for potential security vulnerabilities (e.g., SQL injection, XSS). List each issue with a fix.

Claude outputs a structured list of vulnerabilities and suggested code corrections.

Prompt:

git diff HEAD~1 | claude -p "Review this PR for performance bottlenecks."        

doc: Generate API docstrings and usage examples for 'orderService' class in order_service.ts

Claude updates comments and provides markdown snippet for docs.

Workflow Automation

Example: Automated Code Review

for file in src/*.js; do
  claude -p "Generate JSDoc comments for $file" --output-format text >> docs.md
done        

git diff HEAD~1 | claude -p "Review this PR for performance bottlenecks."

Claude analyzes the patch for speed/efficiency issues and writes a markdown summary.

Example: Batch Documentation

for file in src/*.js; do
  claude -p "Generate JSDoc comments for $file" --output-format text >> docs.md
done        

for file in src/*.js; do claude -p "Generate JSDoc comments for $file" --output-format text >> docs.md done

Auto-generates documentation for each JavaScript file in your src directory.

Advanced Session Management

Example: Multi-step Automation

claude -p "Analyze project structure for modularity" | \
claude -p "Suggest improvements based on analysis" | \
claude -p "Create implementation plan"        

claude -p "Analyze project structure for modularity" | \ claude -p "Suggest improvements based on analysis" | \ claude -p "Create implementation plan"

Chains Claude actions to build a full strategy.

Example: Visual Iteration

Hey Claude! Here's a screenshot of the app homepage. Redesign it for better user engagement using Tailwind CSS, then iterate until the design matches the mock.        

Hey Claude! Here's a screenshot of the app homepage. Redesign it for better user engagement using Tailwind CSS, then iterate until the design matches the mock.

Tips and Best Practices

• Use Markdown checklists for step-by-step code fixes, like linting or migrations.
• Pipe files, logs, or URLs directly into Claude for broader context:

cat error.log | claude -p "Find root cause of repeated crash errors."        

• Update CLAUDE.md with code style, workflow steps, and project notes for smarter agentic behavior.
• Actively course-correct during long sessions; use /clear to refresh Claude’s context window.

With these practical examples, Claude Code becomes a powerful interactive assistant for planning, coding, auditing, documenting, and automating your entire engineering workflow.