Skip to content

Quickstart

Progressive introduction to Claude Code, building understanding from simple interactions to advanced workflows.

1. Simple Prompts

Start with basic questions:

bash
claude
> What files are in this directory?

2. Output Control

Control how Claude responds with /output-style:

> /output-style concise
Fix the authentication bug

Claude gives the fix without explanation. See Output Styles.

3. Planning Mode

For complex tasks, plan before executing with Shift+Tab:

[Press Shift+Tab]
Scaffold a nodejs, typescript REST API with authentication

Review the plan, then approve to proceed.

4. Extended Thinking

For deep problem solving, raise the effort level so the model reasons more before responding:

> /effort high
> Design a multi-tenant SaaS schema

See Extended Thinking for details.

5. Memory

Persist project context across sessions:

> /init
bash
# We use PostgreSQL with Prisma ORM
# Follow REST API conventions

See Memory for details.

6. Delegation

Delegate work to preserve main context:

> Use a subagent to analyze test files and summarize coverage gaps

Subagents work in separate context windows. See Using Subagents.

7. Extensibility

Extend capabilities with MCP servers:

bash
claude mcp add playwright https://mcp.playwright.dev/v1/sse -s user
> Take a screenshot of example.com

See MCP Servers.

Progressive Example

Combining features in a real workflow:

bash
# Initialize context
> /init
TypeScript project with strict mode

# Plan complex work
> Shift+Tab: Design rate limiting system

# Delegate testing
> Use a subagent to generate tests for rate limiter

# Quick concise fix
> /output-style concise
Add JSDoc comments to exported functions

Next Steps

Built with VitePress and powered by Claude Code