CandleKeep
Anthropic

Claude Code: Official Docs

claude-codeskillpluginsagentsanthropic
Pages403
Formatmarkdown
ListedMarch 12, 2026
UpdatedApril 26, 2026
Subscribers58

About

A comprehensive guide to Claude Code covering installation, core concepts, features, automation, integrations, advanced patterns, plugins, security, enterprise, configuration, monitoring, and best practices.

403Chapters
2939Topics
403Pages

Preview

Claude Code: Official Documentation

Comprehensive reference for Claude Code — Anthropic's agentic coding tool. Source: https://github.com/ericbuess/claude-code-docs (community mirror of code.claude.com/docs)



Add to library to read more

Table of Contents

Claude Code: Official Documentation
Overview

How Claude Code Works

Features Overview

Quickstart

You'll be prompted to log in on first use

Setup

Authentication

Desktop Quickstart

Desktop

Interactive Mode

Commands

Cli Reference

Fast Mode

Keybindings

Output Styles

Permissions

Memory

Additional Instructions

Model Config

Start with Opus

Use the opus[1m] or sonnet[1m] alias

Common Workflows

Start Claude in a worktree named "feature-auth"
Creates .claude/worktrees/feature-auth/ with a new branch
Start another session in a separate worktree

Create a worktree with a new branch
Create a worktree with an existing branch
Start Claude in the worktree

Best Practices

Code style
Workflow

One-off queries
Structured output for scripts

Code Review

Hooks

run-tests-async.sh
Read hook input from stdin
Only run tests for source files

Hooks Guide

Mcp

Basic syntax
Real example: Connect to Notion

Basic syntax
Real example: Connect to Asana

Basic syntax

List all configured servers
Get details for a specific server
Remove a server

Add a local-scoped server (default)

Use a custom 5% threshold

Sub Agents

./scripts/validate-readonly-query.sh

Blocks SQL write operations, allows SELECT queries
Read JSON input from stdin
Extract the command field from tool_input using jq

Agent Teams

Skills

Add to deny rules:
Allow only specific skills

Plugins

Plugins Reference

Install to user scope (default)
Install to project scope (shared with team)

Plugin Marketplaces

Discover Plugins

Checkpointing

Scheduled Tasks

Statusline

Tools Reference

Vs Code

Jetbrains

Chrome

Claude Code On The Web

Github Actions

Gitlab Ci Cd

Headless

First request

Remote Control

Amazon Bedrock

Enable Bedrock integration
Optional: Override the region for the small/fast model (Haiku).
Also applies to Bedrock Mantle.
Optional: Override the Bedrock endpoint URL for custom endpoints or gateways

Using inference profile ID
Using application inference profile ARN
Optional: Disable prompt caching if needed

Google Vertex Ai

Set your project ID

Enable Vertex AI integration
Optional: Override the Vertex endpoint URL for custom endpoints or gateways
export ANTHROPIC_VERTEX_BASE_URL=https://aiplatform.googleapis.com
Optional: Disable prompt caching if needed
Optional: Request 1-hour prompt cache TTL instead of the 5-minute default

Microsoft Foundry

Enable Microsoft Foundry integration
Azure resource name (replace {resource} with your resource name)
Or provide the full base URL:

Llm Gateway

Set in environment

~/bin/get-litellm-key.sh
Example: Fetch key from vault
Example: Generate JWT token

Server Managed Settings

Settings

Env Vars

Terminal Config

Network Config

HTTPS proxy (recommended)
HTTP proxy (if HTTPS not available)
Bypass proxy for specific requests - space-separated format
Bypass proxy for specific requests - comma-separated format

Client certificate for authentication
Client private key

Sandboxing

Devcontainer

Security

Data Usage

Zero Data Retention

Legal And Compliance

Costs

Monitoring Usage

1. Enable telemetry
2. Choose exporters (both are optional - configure only what you need)
3. Configure OTLP endpoint (for OTLP exporter)
4. Set authentication (if required)
5. For debugging: reduce export intervals

Console debugging (1-second intervals)
OTLP/gRPC
Prometheus
Multiple exporters
Different endpoints/backends for metrics and logs
Metrics only (no events/logs)

Analytics

Troubleshooting

Reset all user settings and state

macOS (Homebrew)
Windows (winget)
Ubuntu/Debian
Alpine Linux

Third Party Integrations

Slack

Changelog
Changelog
Additional Documentation
Admin Setup

Agent Sdk__agent Loop

Agent Sdk__claude Code Features

Agent Sdk__cost Tracking

Agent Sdk__custom Tools

Agent Sdk__file Checkpointing

Agent Sdk__hooks

Agent Sdk__hosting

Agent Sdk__mcp

Agent Sdk__migration Guide

Before
After
Before

BEFORE (claude-code-sdk)

Agent Sdk__modifying System Prompts

Agent Sdk__observability

Agent Sdk__overview

Agent Sdk__permissions

Agent Sdk__plugins

Agent Sdk__python

Tag a session

Do not load user, project, or local settings from disk
Load only project settings, ignore user and local
Ensure consistent behavior in CI by excluding local settings
Define everything programmatically.
Pass [] to opt out of filesystem setting sources.

Option 1: dict literal (recommended, no import needed)
Option 2: constructor-style (returns a plain dict)

Example conversation:
Turn 1 - You: "Create a file called hello.py"
Turn 1 - Claude: "I'll create a hello.py file for you..."
Turn 2 - You: "What's in that file?"
Turn 2 - Claude: "The hello.py file I just created contains..." (remembers!)
Turn 3 - You: "Add a main function to it"

Agent Sdk__quickstart

Agent Sdk__secure Deployment

Agent Sdk__sessions

Agent Sdk__skills

Check project Skills

Agent Sdk__slash Commands

Agent Sdk__streaming Output

Agent Sdk__streaming Vs Single Mode

Agent Sdk__structured Outputs

Agent Sdk__subagents

Agent Sdk__todo Tracking

Agent Sdk__tool Search

Agent Sdk__typescript V2 Preview

Agent Sdk__typescript

Agent Sdk__user Input

Auto Mode Config

Channels Reference

Testing a plugin you're developing

Channels

Claude Directory

API credentials
Testing Rules

Computer Use

Context Window

Debug Your Config

Desktop Scheduled Tasks

Errors

Fullscreen

Github Enterprise Server

Permission Modes

Platforms

Plugin Dependencies

Routines

Ultraplan

Ultrareview

Voice Dictation

Web Quickstart

Whats New__2026 W13
Week 13 · March 23–27, 2026
Whats New__2026 W14
Week 14 · March 30 – April 3, 2026
Whats New__2026 W15
Week 15 · April 6–10, 2026
Whats New
What's new
Add to Library

Free · Live updates included

58 readers subscribed