Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Initial release: Auto-switch between haiku and sonnet based on task complexity
What This Skill Does
Detects task complexity from message content and switches between a fast model (haiku) and a powerful model (sonnet) automatically. Keyword detection covers Chinese-language terms for analysis, refactoring, architecture, optimization, and debugging. The switch is communicated to the user only when it actually occurs.
Avoids manual model selection by inferring complexity from message content, keeping fast responses for simple tasks and stronger reasoning for complex ones.
When to use it
- Switching to sonnet when asking for a system architecture review
- Staying on haiku for quick single-line code questions
- Batching complex refactoring work on sonnet before switching back
- Auto-detecting Chinese-language analysis requests
- Resetting to the default model after completing a complex session
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: "帮我分析一下这个系统的架构"
- 1Check current model using session_status
- 2Detect keywords in the message: 分析, 架构
- 3Determine that a complex task is detected and current model is haiku
- 4Switch to sonnet using session_status({ model: "kiro-cli" })
- 5Inform user: "切换到 sonnet 模型来处理这个复杂任务..."
Model switched to sonnet; response generated using the more powerful model