openclaw - 💡(How to fix) Fix [Feature]: [Feature Request] Plugin contract should support thinkingOverride to enable dynamic thinking level (low/medium/high/xhigh) per round

Summary

Problem

Currently, OpenClaw's plugin contract does not support thinkingOverride. As a result, it is impossible to dynamically adjust the thinking intensity level (e.g., low / medium / high / xhigh) per round based on the task's requirements.

Current Behavior

The plugin contract uses a fixed or global thinking level. There is no way to override it on a per-round or per-task-context basis.

Desired Behavior

• The plugin contract should support thinkingOverride, allowing each round to independently specify any of the levels: low, medium, high, or xhigh.
• This would enable the plugin to dynamically choose the appropriate thinking budget according to task complexity (e.g., low for simple Q&A, xhigh for complex reasoning).

Proposed Solutions

1. Extend the existing OpenClaw hook contract – add a thinkingOverride parameter or callback to the contract interface.
2. Introduce a dedicated thinking routing hook – separate from the existing contract, solely responsible for per‑round thinking level routing decisions.

Additional Context

This is the only remaining imperfection in OpenClaw regarding dynamic cost/quality trade-offs. Implementing this feature would allow plugins to intelligently adapt to diverse user requests, improving both efficiency and user experience.

Related Notes

• Recorded limitation: thinkingOverride is not supported in the current plugin contract.
• Goal: dynamic selection of low / medium / high / xhigh per round based on the task.

Thank you for considering this feature request!

Problem to solve

/

Proposed solution

/

Alternatives considered

No response

Impact

/

Evidence/examples

No response

Additional information

No response