Start and resume
claudeStart interactive sessionclaude "query"Start interactive session with initial promptclaude -p "query"Query via SDK, then exitcat file | claude -p "query"Process piped contentclaude -cContinue most recent conversation in current directoryclaude -r "<session>" "query"Resume session by ID or nameclaude updateUpdate to latest versionclaude doctorPrint read-only installation and settings diagnostics from the terminal without starting a session, including install health, settings-file validation errors, and Remote Control eligibilityclaude mcpConfigure Model Context Protocol (MCP) serversKeyboard shortcuts
Slash commands
/initInitialize project with a CLAUDE.md guide/clear [name]Start a new conversation with empty context/compact [instructions]Free up context by summarizing the conversation so far/context [all]Visualize current context usage as a colored grid/model [model]Switch the AI model and save it as your default for new sessions/effort [level|auto|status|ultracode [on|off]]Set the effort level: low to xhigh, max, or auto; status prints it/plan [description]Enter plan mode directly from the prompt/resume [session]Resume a conversation by ID or name, or open the session picker/rewindRewind the conversation and/or code to a previous point, or summarize from a selected message/branch [name]Create a branch of the current conversation at this point, so you can try a different direction without losing the conversation as it stands/btw [question]Ask a side question about the current session without adding to the conversation/permissionsManage allow, ask, and deny rules for tool permissions/hooksView hook configurations/mcp [reconnect (<server>|all)|enable|disable [<server>|all]]Manage MCP server connections and OAuth authentication/plugin [subcommand]Manage Claude Code plugins/skillsList available skills/memoryEdit CLAUDE.md files, enable or disable auto memory, and view auto memory entries/review [low|medium|high|xhigh|max|ultra] [--fix] [--comment] [--max-findings n|all|default] [pr#|branch|path]Alias of /code-review: reviews the current diff, or a PR number, branch, or path you pass, such as /review 1234, and takes the same effort levels and flags/security-reviewAnalyze the changes on your current branch for security vulnerabilities/usageShow session cost, plan usage limits, and activity stats/costAlias for /usage/config [key=value ...]Open the Settings interface to adjust theme, model, output style, and other preferences/doctor [prompt-audit [path]]Run a setup checkup that diagnoses installation, settings, extension, and CLAUDE.md problems and proposes fixes that Claude applies after you confirm/statuslineConfigure Claude Code's status linePrompt tricks
@path/to/fileMention a file or folder to pull it into context! npm testShell mode: run a command and show Claude the output/ at startOpen the command and skill menu\ then Enter, or Ctrl+JNew line without sending? on empty promptShow the shortcut help panelultrathinkAsk for deeper reasoning on one turnPaste a screenshotCtrl+V adds images; describe what to fixUseful flags
--modelSets the model for the current session with a model alias such as sonnet, opus, haiku, or fable, or a model's full name--permission-modeBegin in a specified permission mode--allowedToolsTools that execute without prompting for permission, apart from reads from network paths--append-system-promptAppends to the default prompt--output-formatSpecify output format for print mode (options: text, json, stream-json)--max-turnsLimit the number of agentic turns (print mode only)--add-dirAdd additional working directories for Claude to read and edit files--worktreeStart Claude in an isolated git worktree at <repo>/.claude/worktrees/<name>--dangerously-skip-permissionsSkip permission promptsFiles and where they live
CLAUDE.mdProject instructions Claude reads every session. Commit it.CLAUDE.local.mdYour personal project notes. Keep it out of git.~/.claude/CLAUDE.mdInstructions for every project on your machine.claude/settings.jsonShared project settings: permissions, hooks, env, model.claude/settings.local.jsonYour own overrides for this project, gitignored~/.claude/settings.jsonUser settings for every project.claude/agents/*.mdProject subagents (~/.claude/agents/ for yours).claude/skills/<name>/SKILL.mdProject skills and custom slash commands.mcp.jsonProject MCP servers, shared with the team~/.claude.jsonYour MCP servers, OAuth sessions and per-project state~/.claude/keybindings.jsonCustom keyboard shortcutsGenerate these files instead of writing them by hand: CLAUDE.md, settings.json, subagents, .mcp.json.
Frequently asked questions
What are the most important Claude Code commands?
For daily work: /init to create a CLAUDE.md, /clear between unrelated tasks, /compact when context fills up, /model to switch models, /plan to plan before editing, /resume to reopen a session and /rewind to undo. claude -p runs Claude Code once from a script.
How do I undo what Claude Code just did?
Press Esc twice on an empty prompt, or run /rewind, to go back to an earlier checkpoint. You can restore the code, the conversation or both.
How do I give Claude Code project instructions?
Put them in CLAUDE.md at the root of the repo. Run /init to generate a first draft, or use the CLAUDE.md generator on this site, and keep it short: build and test commands, conventions and the gotchas Claude would not guess.
Is there a printable Claude Code cheat sheet?
Yes, this page. Use your browser's print or Save as PDF; the layout switches to a compact two-column sheet when printed.
Sources: Commands, CLI reference, Interactive mode. Claude Code Agents is an independent site, not affiliated with Anthropic.