Progress will not be saved. This browser is blocking storage (often private browsing). You can still learn, but closing the tab loses your progress.
CLI commands and flags
--add-dirGrants Claude file read and write permissions in extra folders.--advisorActivates the server advisor feature for the session using a chosen model alias or identifier.--agentSelects a specific agent implementation for the current session.--agentsPasses custom subagent specifications dynamically using raw JSON or a JSON file path.--allow-dangerously-skip-permissionsPermits cycling into the bypassPermissions option during Shift+Tab mode changes.--allowedToolsSpecifies tools that can run automatically without prompting for authorization.--append-subagent-system-promptAdds extra instructions to the end of system prompts across all subagents in non-interactive runs.--append-subagent-system-prompt-fileReads text from a file to append to subagent system prompts during non-interactive runs.--append-system-promptAdds user-defined instructions to the conclusion of the standard system prompt.--append-system-prompt-fileLoads a file to attach its contents to the default system prompt.--autocompactConfigures the auto-compact threshold for the active session without altering stored settings.--ax-screen-readerSwitches output to plain text without decorative formatting to aid screen reader software.--bareDisables loading project configuration, hooks, plugins, and MCP servers for faster startup.--betasIncludes custom beta feature headers when dispatching API calls.--bgSpawns the agent as a background task and immediately returns control to the shell.--channelsListens for MCP server channel notifications from specified plugins during the session.--chromeTurns on Chrome integration to perform automated web actions and tests.--cloudStarts a new cloud session with a prompt or pushes a message to an existing session.--continueRestores the latest conversation found in the current working directory.--correlation-idProvides an identifier to the spawn-runner hook alongside the environment option.--dangerously-load-development-channelsPermits unapproved channels for local development after interactive confirmation.--dangerously-skip-permissionsBypasses all permission confirmations throughout the session.--debugActivates debugging output with optional category inclusions or exclusions.--debug-fileDirects debug output to a designated file and turns on debug mode.--desktopLaunches the Claude Desktop application in this directory without running in terminal.--disable-slash-commandsDeactivates all skills and slash commands for the active session.--disallowedToolsSets tool restriction rules by hiding tools or blocking specific invocations.--effortSpecifies the reasoning effort level applied to the primary conversation model.--enable-auto-modeIndicates an obsolete flag replaced by auto permission mode.--environmentInitializes a cloud session targeting the designated self-hosted runner pool.--exclude-dynamic-system-prompt-sectionsShifts per-user context to the initial user prompt to improve prompt caching.--execExecutes a shell command as a background PTY process instead of running Claude.--fallback-modelConfigures ordered alternative models to handle outages of the primary model.--fork-sessionGenerates a fresh session ID when reopening an existing conversation.--forward-subagent-textStreams text and reasoning outputs from subagents into the main stream.--from-prFilters the session selector to conversations linked with a given pull request.--ideConnects automatically to the IDE when a single valid instance is present.--include-hook-eventsStreams hook lifecycle progress and responses in JSON streaming output.--include-partial-messagesOutputs streaming events incrementally when running in print mode with stream-json output.--initTriggers init Setup hooks prior to starting the session in print mode.--init-onlyExecutes Setup and SessionStart hooks then immediately exits.--input-formatChooses the input format for print mode between text and stream-json.--json-schemaEnforces output validation against a specified JSON Schema once agent execution finishes in print mode.--maintenanceTriggers Setup hooks configured with the maintenance matcher prior to starting a print mode session.--max-budget-usdHalts print mode execution when client-side estimated API expenses reach the given dollar threshold.--max-turnsRestricts the total number of agent interaction turns allowed in print mode.--mcp-configLoads MCP servers from provided JSON configuration files or raw JSON strings.--metadataApplies key-value metadata tags to an environment-based session for matching orchestrator hook labels.--modelSelects the model for the session using a full identifier or an alias like sonnet or opus.--nameAssigns a custom title to the session for display in resume lists and the terminal bar.--no-chromeTurns off Chrome browser integration for the active session.--no-session-persistencePrevents print mode sessions from being saved to disk or resumed later.--output-formatSets the format of the output in print mode to text, json, or stream-json.--permission-modeSelects the initial permission handling mode for the session, such as auto, plan, or manual.--permission-prompt-toolDesignates an MCP tool to process authorization requests in non-interactive mode.--permission-promptsConfigures how permission prompts are resolved in print mode, allowing host forwarding or denial via none.--plugin-dirLoads one or more plugins for the session from directory paths or zip files.--plugin-urlFetches and applies remote plugin zip archives from one or more URLs for the session.--printRuns Claude Code non-interactively and prints the response directly to standard output.--prompt-suggestionsOutputs suggested follow-up user prompts after turns when streaming JSON in print mode.--refSpecifies a Git reference branch to check out when provisioning an environment session.--remoteServes as a deprecated alias for the --cloud flag.--remote-controlLaunches an interactive session manageable from claude.ai or the Claude app, with an optional name for the session.--remote-control-session-name-prefixSets a prefix for automatically generated Remote Control session identifiers.--replay-user-messagesEchoes user input messages from stdin back onto stdout when using stream-json I/O.--restrictedLaunches Claude Code in a restricted mode that disables command execution tools and limits file access to working directories.--resumeRestores a prior session using its ID, name, or transcript path, or opens an interactive list to choose one.--safe-modeStarts Claude Code with all customizations turned off to help debug broken configurations.--session-idAssigns a specific UUID to be used as the conversation session ID.--setting-sourcesSpecifies a comma-separated list of configuration sources to load among user, project, and local.--settingsOverrides settings using a JSON file under 2 MiB or an inline JSON string for the current session.--strict-mcp-configIgnores existing MCP configurations from standard files or plugins and loads only the ones specified via --mcp-config.--system-promptSubstitutes the entire default system prompt with the provided custom text.--system-prompt-fileReads system prompt content from a file and replaces the default prompt.--system-prompt-snapshotForces Claude Code to reconstruct the system prompt on each request instead of reusing the initial prompt snapshot.--teammate-modeDefines how agent teammates are shown on screen, overriding the teammateMode setting.--teleportBrings an active cloud session back into the local terminal.--tmuxLaunches a tmux session alongside a git worktree, using iTerm2 panes when supported.--toolsControls which built-in tools Claude is allowed to use during the session.--verboseEnables detailed logging to display the complete turn-by-turn output.--versionDisplays the installed version number of Claude Code.--worktreeLaunches Claude inside an isolated git worktree, optionally checked out from a pull request or merge request.claudeLaunches an interactive working session.claude "query"Opens an interactive session preloaded with a starting prompt.claude -cResumes the previous chat held within the current folder.claude -c -p "query"Carries on a previous interaction via SDK.claude -p "query"Sends a prompt via SDK and terminates upon completion.claude -r "<session>" "query"Reopens a specific session using its identifier or title.claude agentsLaunches an interface in the terminal to inspect and spawn concurrent background sessions.claude attach <id|name>Binds the active terminal to a background task by identifier or name.claude auth loginAuthenticates to Anthropic with options for email, SSO, or Console billing.claude auth logoutSigns out of the current Anthropic account.claude auth statusOutputs login details as JSON or formatted text via --text.claude auto-mode defaultsOutputs the built-in classification rules for auto mode in JSON format.claude auto-mode resetDeletes the autoMode section from user preferences to reinstate initial defaults.claude daemon logsStreams live output from ~/.claude/daemon.log until interrupted.claude daemon runExecutes the background supervisor directly in the active terminal to display logs.claude daemon statusShows supervisor health, current version, socket path, and active worker count.claude daemon stop --anyHalts the supervisor and its active tasks, optionally preserving workers via --keep-workers.claude doctorEvaluates setup health, configuration syntax, and Remote Control support without starting a session.claude gatewayLaunches a self-hosted gateway server using a mandatory --config gateway.yaml.claude import [source]Launches a session running /import to migrate configurations from external coding agents.claude install [version]Installs or replaces the native executable using a specific version, stable, or latest.claude logs <id|name>Displays the latest log entries generated by a background session.claude mcpManages configuration settings for MCP servers.claude mcp login <name>Executes OAuth authentication for a specified MCP server directly from the command line.claude mcp logout <name>Removes saved OAuth authentication credentials for a specified MCP server.claude pluginProvides commands to handle Claude Code plugins.claude purge [path]Erases all local state, history, logs, and configuration records for a project.claude remote-controlLaunches a server mode instance enabling claude.ai or the mobile app to operate Claude Code.claude respawn <id>Relaunches an active or inactive background session while preserving its conversation history.claude rm <id>Deletes a background session from tracking while keeping transcripts stored locally.claude self-hosted-runnerRegisters the machine to host cloud-initiated Claude Code sessions on private infrastructure.claude setup-tokenOutputs an unsaved, persistent OAuth token directly to the terminal for automation use.claude stop <id>Terminates an active background session.claude ultrareview [target]Performs an automated review, outputs results directly to stdout, and can submit comments to a PR.claude updateUpgrades the application to the most recent release.
Nothing matches.