OpenClaw Slash Commands
Slash commands are typed directly into the OpenClaw chat interface to control sessions, models, agents, and platform behavior. Commands marked owner-only require elevated permissions.
Quick Reference
| Command | Description |
|---|---|
/help |
Shows available slash commands |
/commands |
Lists all enabled slash commands |
/status |
Shows current status, quotas, and provider health |
/context [list\|detail\|json] |
Shows context usage details |
/whoami |
Shows your sender ID (alias: /id) |
/reset |
Resets conversation (alias: /new) |
/new [model] |
Starts fresh session with optional model |
/compact [instructions] |
Compacts conversation context |
/session idle <duration\|off> |
Manages inactivity auto-unfocus |
/session max-age <duration\|off> |
Manages session max-age auto-unfocus |
/export-session [path] |
Exports current session to HTML |
/stop |
Stops current operation |
/restart |
Restarts the service |
/model <name> |
Switches active model (alias: /models) |
/think <level> |
Sets thinking depth (aliases: /thinking, /t) |
/fast [status\|on\|off] |
Toggles fast mode |
/verbose [on\|full\|off] |
Controls verbosity (alias: /v) |
/reasoning [on\|off\|stream] |
Shows reasoning steps (alias: /reason) |
/usage [off\|tokens\|full\|cost] |
Controls usage footer display |
/tts [...] |
Controls text-to-speech settings |
/skill <name> [input] |
Runs a specific skill by name |
/subagents [...] |
Manages sub-agent runs |
/agents |
Lists thread-bound agents |
/kill <id\|#\|all> |
Aborts running sub-agents |
/steer <id\|#> <message> |
Steers running sub-agent (alias: /tell) |
/btw <question> |
Asks ephemeral side question without affecting context |
/acp [...] |
Controls ACP runtime sessions |
/exec [...] |
Configures execution security settings |
/elevated [on\|off\|ask\|full] |
Sets elevated permissions (alias: /elev) |
/bash <command> |
Executes host command (host-only) |
! <command> |
Host command alternative (host-only) |
!poll |
Checks bash job status |
!stop |
Stops running bash job |
/allowlist |
Lists/adds/removes allowlist entries |
/approve <id> [...] |
Resolves execution approval prompts |
/config [show\|get\|set\|unset] |
Persists config to disk (owner-only) |
/mcp [show\|get\|set\|unset] |
Manages MCP server config (owner-only) |
/plugins [...] |
Toggles plugin enablement (owner-only) |
/debug [...] |
Runtime debug overrides (owner-only) |
/send [on\|off\|inherit] |
Controls send permissions (owner-only) |
/queue <mode> |
Configures queue behavior |
/focus <target> |
Binds thread to session/subagent (Discord) |
/unfocus |
Removes current thread binding (Discord) |
/dock-telegram |
Switches replies to Telegram |
/dock-discord |
Switches replies to Discord |
/dock-slack |
Switches replies to Slack |
/activation [mention\|always] |
Controls activation mode (groups only) |
Aliases & Shortcuts
Frequently used commands with their shorter aliases for faster typing.
| Command | Alias | Description |
|---|---|---|
/whoami |
/id |
Shows your sender ID |
/reset |
/new |
Resets the conversation |
/model <name> |
/models |
Switches the active model |
/think <level> |
/thinking, /t |
Sets thinking depth |
/verbose [on\|full\|off] |
/v |
Controls response verbosity |
/reasoning [on\|off\|stream] |
/reason |
Shows reasoning steps |
/steer <id\|#> <message> |
/tell |
Steers a running sub-agent |
/elevated [on\|off\|ask\|full] |
/elev |
Sets elevated permissions |
! <command> |
/bash <command> |
Executes a host command |
Help & Info
Quick access to status, context usage, and your current identity.
/help
Shows all available slash commands in the current session.
/commands
Lists every enabled slash command, including those added by plugins.
/status
Displays current system status — active quotas, provider health, and session state.
/context [list | detail | json]
Shows how much context has been consumed and what is currently loaded. Useful for managing long sessions.
/whoami
Returns your sender ID. Alias: /id
Session Management
Commands for controlling conversation lifecycle, compaction, and session expiry.
/reset
Resets the current conversation to a clean state. Alias: /new
/new [model]
Starts a fresh session, optionally switching to a different model on launch.
/compact [instructions]
Compacts the current conversation context to reduce token usage while preserving key information. Accepts optional instructions to guide what is kept.
/session idle <duration | off>
Sets an inactivity timeout after which the session auto-unfocuses. Pass off to disable.
/session max-age <duration | off>
Caps the maximum age of a session before it auto-unfocuses, regardless of activity. Pass off to disable.
/export-session [path]
Exports the current session as an HTML file. Optionally specify a file path.
/stop
Stops the currently running operation immediately.
/restart
Restarts the OpenClaw service.
Model & Output Control
Tune model behavior, verbosity, and output formatting on the fly.
/model <name>
Switches the active model mid-session. Alias: /models
/think <level>
Sets the thinking depth for the model's reasoning process. Aliases: /thinking, /t
/fast [status | on | off]
Toggles fast mode for quicker responses. Use status to check the current state.
/verbose [on | full | off]
Controls how much detail appears in responses. Alias: /v
/reasoning [on | off | stream]
Shows or hides the model's reasoning steps inline. Use stream to display them as they are generated. Alias: /reason
/usage [off | tokens | full | cost]
Controls what appears in the usage footer — token counts, cost estimates, or nothing.
/tts [off | always | inbound | tagged | status | provider | limit | summary | audio]
Configures text-to-speech behavior — when it fires, which provider is used, and audio settings.
Sub-agents & Skills
Manage background agents and run skills from the interface.
/skill <name> [input]
Runs a named skill directly, with optional input passed as an argument.
/subagents [list | kill | log | info | send | steer | spawn]
Full lifecycle management for sub-agent runs — inspect, control, or launch agents.
/agents
Lists all agents currently bound to the active thread.
/kill <id | # | all>
Aborts one or all running sub-agents immediately.
/steer <id | #> <message>
Sends a steering message to redirect a running sub-agent. Alias: /tell
/btw <question>
Asks an ephemeral side question that does not affect the main session context — useful for quick lookups mid-task.
ACP & Execution
Control ACP runtime sessions and command execution settings.
/acp [spawn | cancel | steer | close | status | set-mode | set | cwd | permissions | timeout | model | reset-options | doctor | install | sessions]
Full control over ACP (Agent Control Protocol) runtime sessions — spawn, monitor, configure, and terminate.
/exec [...]
Configures execution security settings for shell and tool calls.
/elevated [on | off | ask | full]
Sets the elevated permissions level for the current session. Alias: /elev
/bash <command>
Executes a command on the host system directly. Host-only.
! <command>
Alternative syntax for host command execution. Host-only.
!poll
Checks the status of a background bash job.
!stop
Stops a currently running background bash job.
Permissions & Config
Owner-only commands for managing access control, configuration persistence, and plugins.
/allowlist
Lists, adds, or removes entries from the execution allowlist.
/approve <id> [allow-once | allow-always | deny]
Resolves a pending execution approval prompt — grant once, grant permanently, or deny.
/config [show | get | set | unset]
Reads or writes persistent configuration values to disk. Owner-only.
/mcp [show | get | set | unset]
Manages MCP server configuration. Owner-only.
/plugins [list | show | get | enable | disable]
Enables or disables installed plugins. Owner-only.
/debug [show | set | unset | reset]
Applies or removes runtime debug overrides without restarting. Owner-only.
/send [on | off | inherit]
Controls whether send permissions are active for the current context. Owner-only.
/queue <mode>
Configures how incoming messages are queued during busy periods.
Platform & Focus
Commands for binding sessions to specific platforms and switching reply destinations.
/focus <target>
Binds the current thread to a specific session or sub-agent. Discord-specific.
/unfocus
Removes the current thread binding, returning to the default session. Discord-specific.
/dock-telegram
Switches all replies to Telegram.
/dock-discord
Switches all replies to Discord.
/dock-slack
Switches all replies to Slack.
/activation [mention | always]
Controls when the bot activates in group contexts — on mention only, or always. Groups only.