claude-documentation

Claude Code Documentation Index

Fetch from this index before implementing any Claude Code feature — do not guess at behavior. One URL per topic — read the most specific one first.

Discovery

• https://code.claude.com/docs/llms.txt — Complete index of all doc pages; fetch when you need a URL not listed here.

Core Reference

• https://code.claude.com/docs/en/cli-reference — CLI flags, startup arguments, --worktree, --settings, --model, non-interactive mode (-p).
• https://code.claude.com/docs/en/settings — All configurable settings, settings file locations and scopes (managed > local > project > user), worktree settings.
• https://code.claude.com/docs/en/env-vars — Environment variables; use when a behavior can be toggled by env var.
• https://code.claude.com/docs/en/tools-reference — All built-in tools (Read, Edit, Bash, WebFetch, Agent, etc.), permission rule syntax per tool, and tool behavior details.

Extension Points

• https://code.claude.com/docs/en/skills — SKILL.md format, all frontmatter fields (allowed-tools, user-invocable, disable-model-invocation, context, agent), dynamic context injection, supporting files.
• https://code.claude.com/docs/en/hooks — Hook lifecycle reference: events (PreToolUse, PostToolUse, Stop, SessionStart, statusLine, WorktreeCreate…), handler types (command/HTTP/MCP/prompt/agent), full input/output JSON schemas.
• https://code.claude.com/docs/en/statusline — Status line setup: configuration, available data fields (model, cost, context %, worktree, effort, vim mode, rate limits…), example scripts.
• https://code.claude.com/docs/en/plugins-reference — Plugin component schemas for skills, agents, hooks, MCP servers, monitors, themes; CLI commands for plugin management.

Permissions & Security

• https://code.claude.com/docs/en/permissions — Permission rules (allow/ask/deny), all permission modes, Bash/WebFetch/Read/Edit rule syntax, working directories, managed settings.
• https://code.claude.com/docs/en/model-config — Model aliases (opus, sonnet, opusplan), effort levels, extended thinking, extended context, model-switching via flag/env/settings.

Session & Interaction

• https://code.claude.com/docs/en/commands — All built-in /commands and bundled skills (/loop, /batch, /code-review, /debug) available in a session.
• https://code.claude.com/docs/en/interactive-mode — Keyboard shortcuts, vim editor mode, shell mode (! prefix), task list, PR review status, session recap.
• https://code.claude.com/docs/en/keybindings — Customizing keybindings: contexts, all action names, chord syntax, unbinding defaults.
• https://code.claude.com/docs/en/terminal-config — Terminal setup: Shift+Enter, Option-as-Meta on macOS, tmux passthrough, fullscreen rendering, vim mode.

Parallel Workflows

• https://code.claude.com/docs/en/worktrees — --worktree flag, subagent isolation, .worktreeinclude for copying gitignored files, WorktreeCreate/WorktreeRemove hooks, cleanup.

Context & Memory

• https://code.claude.com/docs/en/memory — CLAUDE.md files (project/user/managed), auto memory, .claude/rules/ with path-scoped frontmatter.
• https://code.claude.com/docs/en/glossary — Definitions: agentic loop, compaction, hook, skill, subagent, worktree isolation, session, tool, permission mode, etc.