Verbose in chat despite changes

[molten fractal]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[torn patio]

That kind of output (“The user is frustrated… Let me look at the screenshot…”) usually isn’t OpenClaw’s /verbose mode — it’s the model printing its own “thinking/planning” text.

With Moonshot/Kimi, this most often happens when native thinking is enabled (either because you’re on a …thinking… model, or because the session is inheriting a non-off thinking default).

1) First: what version + what exact model?

Please paste:

• openclaw --version
• in the noisy chat: /model status (or tell me the exact moonshot/... model id)

2) Quick fix in-chat (no config changes)

In the same chat/session where it’s being verbose, send these as standalone messages:

/think off
/reasoning off
/verbose off

Then send a normal test message. If this fixes it, the root cause was “thinking” being enabled via inherit.

Docs: https://docs.openclaw.ai/tools/thinking

3) If you’re accidentally on a “thinking” Kimi model

If /model status shows something like:

• moonshot/kimi-k2-thinking or moonshot/kimi-k2-thinking-turbo

…switch to the non-thinking one:

• moonshot/kimi-k2.5

Docs: https://docs.openclaw.ai/providers/moonshot

4) Make it permanent (config)

If you want to force-disable Moonshot thinking regardless of session inheritance, set this in ~/.openclaw/openclaw.json:

{
  agents: {
    defaults: {
      models: {
        "moonshot/kimi-k2.5": {
          params: {
            thinking: { type: "disabled" },
          },
        },
      },
      model: {
        thinkingDefault: "off",
      },
    },
  },

}

Then restart the gateway (openclaw gateway restart).

If you paste openclaw --version + /model status, I can tell you which of the above applies immediately.

[plain warren]

My discord glitched and I posted multiple. Please delete thread