Chat commands, aborting tasks, and stopping a run

Most internal/tool messages only appear when verbose, trace, or reasoning is enabled for that session.

Fix in the chat where you see it:

/verbose off/trace off/reasoning off

Still noisy: check session settings in the Control UI and set verbose to inherit; confirm you are not using a bot profile with verboseDefault: "on" in config.

Docs: Thinking and verbose, Security.

Send any of these as a standalone message (no slash) to trigger an abort: stop, stop action, stop current action, stop run, stop current run, stop agent, stop the agent, stop openclaw, openclaw stop, stop don't do anything, stop do not do anything, stop doing anything, do not do that, please stop, stop please, abort, esc, exit, interrupt, halt. Common non-English triggers (French, German, Spanish, Chinese, Japanese, Hindi, Arabic, Russian) also work.

For background processes started by the exec tool, ask the agent to run:

process action:kill sessionId:XXX

Most slash commands must be sent as a standalone message starting with /, but a few shortcuts (like /status) also work inline for allowlisted senders. See Slash commands.

OpenClaw blocks cross-provider messaging by default. If a tool call is bound to Telegram, it will not send to Discord unless you explicitly allow it - and this takes effect immediately, no gateway restart needed:

{  tools: {    message: {      crossContext: {        allowAcrossProviders: true,        marker: { enabled: true, prefix: "[from {channel}] " },      },    },  },}

Mid-run prompts are steered into the active run by default. Use /queue to choose active-run behavior:

• steer (default) - guide the active run at the next model boundary.
• followup - queue messages and run them one at a time after the current run ends.
• collect - queue compatible messages and reply once after the current run ends.
• interrupt - abort the current run and start fresh.

Add options to queued modes like debounce:0.5s cap:25 drop:summarize. See Command queue and Steering queue.