dvcrn Model Switcher
Automatically detect task complexity and switch to the appropriate model.
Model Configuration
Based on your setup:
- Fast model (haiku):
custom-kiro-cli-vipdump-eu-org/claude-haiku-4-5- For simple, quick tasks - Powerful model (sonnet):
custom-kiro-cli-vipdump-eu-org/claude-sonnet-4-5- For complex analysis
Trigger Keywords (Chinese)
Switch to sonnet when message contains:
- 分析、深入分析、详细分析
- 重构、代码重构
- 架构、系统架构、设计架构
- 设计、系统设计
- 优化、性能优化
- 复杂、复杂问题
- 调试、深度调试
- 评估、技术评估
Workflow
- Check current model using
session_status - Detect keywords in user message
- Switch if needed:
- Complex task detected + currently on haiku → switch to sonnet
- Simple task + currently on sonnet → optionally switch back to haiku
- Inform user briefly about the switch
Implementation
Use session_status tool with model parameter:
// Switch to sonnet for complex tasks
session_status({ model: "kiro-cli" })
// Switch back to haiku for simple tasks
session_status({ model: "haiku" })
// Reset to default
session_status({ model: "default" })
Example Detection Logic
User message: "帮我分析一下这个系统的架构"
→ Contains: "分析", "架构"
→ Action: Switch to sonnet if not already on it
→ Response: "🔄 切换到 sonnet 模型来处理这个复杂任务..."
Notes
- Be smart about switching - don't switch for every message
- Batch related complex tasks on sonnet before switching back
- Inform user only on actual switches, not when already on correct model
- Consider context: if already discussing complex topic, stay on sonnet