Guides/Prompt Shortcuts

32 Claude Code Prompt Shortcuts + Complete Command Reference

Every built-in command, keyboard shortcut, and prompt pattern you need to get maximum output from Claude Code. Copy, paste, and combine for instant results.

PromptingCommandsReferenceEssential

Built-in Commands

These commands are available in every Claude Code session. Type them directly into the prompt.

Session Commands

/helpShow all available commands

/clearClear conversation history and start fresh

/compactCompress conversation to free up context window

/costShow token usage and estimated cost for this session

/statusShow current working directory, git branch, and session info

/quitEnd the session

Model & Mode Commands

/modelSwitch between models (Opus, Sonnet, Haiku)

/fastToggle fast mode (faster output, same model)

/thinkEnable extended thinking for complex reasoning

/permissionsView or change tool permissions

Agent Commands

/agentsLaunch a background agent that works independently

/bugReport a bug to Anthropic

File & Context References

@file.tsReference a specific file in your prompt

@folder/Reference an entire folder

@urlFetch and include web content

!commandRun a shell command inline

Custom Commands

Create .claude/commands/name.md files to define your own slash commands. Examples: /start, /sync, /wrap-up, /review, /deploy

Keyboard Shortcuts

Speed up your workflow with these keyboard shortcuts available in the Claude Code terminal.

EnterSend message

Shift+EnterNew line (don't send)

EscapeCancel current generation

Ctrl+CInterrupt current operation

Ctrl+LClear screen

Up ArrowRecall previous message

TabAuto-complete file paths

32 Prompt Patterns

Add these prefixes at the start of your prompt to change how Claude responds. Example: ELI5: How does React Server Components work?

Explanation5 patterns

ELI5:Explain like I'm 5 — simple, no jargon

ELI5: What is a webhook?

TLDR:Summarize in 2-3 sentences max

TLDR: What does this function do?

STEP BY STEP:Break down into numbered steps

STEP BY STEP: How to set up auth

FIRST PRINCIPLES:Rebuild from fundamental basics

FIRST PRINCIPLES: Why do we need state management?

CHAIN OF THOUGHT:Show intermediate reasoning

CHAIN OF THOUGHT: Debug this error

Format5 patterns

CHECKLIST:Turn response into a checklist

CHECKLIST: Pre-deploy verification

TABLE:Format as a comparison table

TABLE: Compare React vs Vue vs Svelte

JSON:Output as structured JSON

JSON: List all API endpoints in this project

SCHEMA:Generate a data model or outline

SCHEMA: Design the database for a todo app

FORMAT AS [X]:Force a specific format

FORMAT AS YAML: Show the config structure

Role4 patterns

ACT AS [role]:Adopt a specific persona

ACT AS security auditor: Review this auth code

DEV MODE:Raw technical developer response

DEV MODE: What's wrong with this regex?

PM MODE:Project management perspective

PM MODE: Break this feature into tasks

ARCHITECT:System design perspective

ARCHITECT: How should we structure this API?

Analysis5 patterns

SWOT:Strengths / Weaknesses / Opportunities / Threats

SWOT: Using SQLite vs Postgres for this project

COMPARE:Side-by-side comparison

COMPARE: JWT vs session-based auth

PITFALLS:Identify traps and common mistakes

PITFALLS: Deploying Next.js to Vercel

TRADEOFFS:Analyze pros and cons of each approach

TRADEOFFS: Monorepo vs polyrepo

METRICS:Express with measures and indicators

METRICS: How to measure API performance

Thinking5 patterns

MULTI-PERSPECTIVE:Show several viewpoints

MULTI-PERSPECTIVE: Should we use microservices?

DELIBERATE:Slow, thoughtful reasoning

DELIBERATE: Design the error handling strategy

REFLECTIVE:AI reflects on its own answer

REFLECTIVE: Is this the best approach for caching?

EVAL-SELF:Critical self-evaluation

EVAL-SELF: Rate this code 1-10 and explain why

BIAS CHECK:Identify potential biases

BIAS CHECK: Am I over-engineering this?

Control7 patterns

BRIEFLY:Force a very short answer

BRIEFLY: What does useEffect do?

AUDIENCE [X]:Adapt to a specific audience

AUDIENCE beginner: Explain Git branching

TONE [X]:Change the tone

TONE casual: Explain this error message

JARGON:Use technical vocabulary

JARGON: Describe the rendering pipeline

NO AUTOPILOT:Force careful, non-generic response

NO AUTOPILOT: Review this PR for real issues

GUARDRAIL [X]:Set strict boundaries

GUARDRAIL no new dependencies: Fix this bug

CONTEXT:Keep multiple context layers

CONTEXT: We're building a SaaS, using Next.js 16, deploying to Vercel

Power Combinations

Combine multiple patterns in a single prompt for maximum effectiveness. Stack prefixes, separated by commas, before your actual request.

ACT AS security auditor, CHECKLIST:

Review this authentication flow for OWASP Top 10 vulnerabilities

ARCHITECT, COMPARE, TABLE:

Evaluate 3 approaches for real-time notifications (WebSockets vs SSE vs Polling)

STEP BY STEP, GUARDRAIL no breaking changes:

Migrate from Pages Router to App Router

DEV MODE, PITFALLS, BRIEFLY:

What could go wrong with this database migration?

FIRST PRINCIPLES, DELIBERATE:

Should we build our own auth or use a third-party service?

Claude Code-Specific Prompts

These prompts leverage Claude Code's built-in tools — Read, Write, Edit, Grep, Glob, and Bash — to interact directly with your codebase and terminal.

Read [file] and explain what it does

Read tool

Find all files that import [module]

Grep / Glob

Run the tests and fix any failures

Bash + Edit

Create a new component for [X]

Write tool

Refactor [file] to use [pattern]

Read + Edit

Search the codebase for security issues

Grep + analysis

What's the git status?

Bash (git)

Create a PR for these changes

Bash (gh CLI)

Build this project and show me the errors

Bash (build)

Install [package] and set it up

Bash (npm/pip)

Prompt Templates for Common Tasks

Copy these templates and fill in the brackets for common development workflows.

Bug Fix Template

I'm seeing this error: [paste error]

It happens when: [describe when]

Expected behavior: [what should happen]

Fix it and explain what was wrong.

Feature Request Template

Build [feature name].

Requirements:

1. [requirement 1]

2. [requirement 2]

3. [requirement 3]

Use [technology/pattern].

Write tests for it.

Code Review Template

Review [file/PR] for:

- Bugs and edge cases

- Security vulnerabilities

- Performance issues

- TypeScript type safety

- Code style and readability

Rate each area 1-10.

Refactor Template

Refactor [file] to:

- [improvement 1]

- [improvement 2]

GUARDRAIL: Don't change the public API.

GUARDRAIL: All existing tests must still pass.

Get All 32 Shortcuts + 143 Skills

The AI Brain Pro package includes this complete shortcut reference, 143 pre-built skills, 130 agents, and a 7-tier memory system that makes Claude Code work like a senior engineer.

Get AI Brain Pro — $97 Browse All Guides