About this project

This repository is a community-maintained reference guide for Anthropic's Claude Code command-line tool. It is documentation rather than executable software, and it explicitly points readers to the official Claude Code documentation as the authoritative source, noting that commands and provider model mappings change quickly. What the guide covers: - Getting started: quick-start installation paths (native installer scripts for macOS/Linux/WSL, Windows PowerShell and CMD, Homebrew, WinGet, and npm with Node.js 22+), version verification, and system requirements (macOS 13+, Windows 10 1809+/Server 2019+, Ubuntu 20.04+, Debian 10+, Alpine 3.19+, 4 GB+ RAM, x64/ARM64). - Initial setup and authentication: browser sign-in via `claude auth login`, Console/API billing mode, and guidance to inject credentials from an OS key store or secret manager rather than committing them. It warns that persistent API keys or auth tokens select API/provider authentication and can disable subscription-only features. - Configuration and environment: a catalog of environment variables for authentication and routing, model selection and provider alias overrides, third-party providers (Bedrock, Vertex, Foundry), timeouts and output budgets, session/context/agent limits, accessibility, feature flags, proxy settings, and privacy/telemetry toggles. It notes that settings.json values override shell values and that some variables are presence-based. - Configuration files: a table of memory types and locations, including enterprise policy, project memory (CLAUDE.md), user memory, local project memory, and modular `.claude/rules/` files, plus auto-memory behavior and the `@path` import mechanism. - Commands and usage: a snapshot table of slash commands (for example `/config`, `/agents`, `/mcp`, `/memory`, `/permissions`, `/plan`, `/code-review`, `/rewind`, `/sandbox`, `/usage`, `/workflows`) and a command-line flags table. - Advanced features: sections on thinking keywords, effort levels, advisor tool, fast mode, auto mode, plan mode, background tasks, workflows and scheduling, remote sessions, browser and IDE integrations, sandbox mode, LSP tool, subagents, agent teams, skills, plugin system, worktree isolation, native installer, authentication and agent-management CLIs, remote control, managed settings, model updates, theming, code review, insights, MCP integration, and the hooks system. - Security and permissions: dangerous mode, permission modes, and security best practices. - Automation and integration: scripting with Claude Code, automated PR review with inline comments, and issue triage with suggested labels and severity. - Help and troubleshooting: installation/Node.js issues, MCP issues, best practices, and monitoring/alerting. - Third-party integrations: provider setup examples and a DeepSeek integration section. The README also links to related community resources, including a Discord integration, security agent SKILL.md files, an agent-skills authoring project, a large collection of agent skills, no-cost AI resources, Mermaid templates, and a Discord MCP agent. The repository is MIT licensed and marked active. Because it is a guide, its value is in consolidating setup steps, configuration references, and workflow patterns in one place; readers should verify current command names and flags against the official documentation, as the guide itself advises.