OpenClaw · Skill
Dynamic Model Selector
This skill analyzes user queries to recommend the optimal AI model from available GitHub Copilot options, balancing performance, cost, and task requirements.
Install
Start with the primary install command. Alternate entrypoints are included below for ClawHub and OpenClaw CLI users.
Primary command
clawhub install mpelissari/dynamic-model-selectorClawHub installer
npx clawhub@latest install mpelissari/dynamic-model-selectorOpenClaw CLI
openclaw skills install mpelissari/dynamic-model-selectorDirect OpenClaw install
openclaw install mpelissari/dynamic-model-selectorWhat this skill does
This skill analyzes user queries to recommend the optimal AI model from available GitHub Copilot options, balancing performance, cost, and task requirements.
Why it matters
Instead of manually picking a model each time, the classification script matches task complexity to the right model, avoiding overpaying for simple queries or underserving complex ones.
Typical use cases
- Routing simple chat queries to free models to avoid unnecessary cost
- Selecting gpt-4o for complex analysis tasks automatically
- Picking code-optimized models for code generation requests
- Comparing model options before starting a long reasoning task
- Keeping costs low on high-volume automated workflows
Source instructions
Dynamic Model Selector
Overview
This skill analyzes user queries to recommend the optimal AI model from available GitHub Copilot options, balancing performance, cost, and task requirements.
How to Use
- Provide the user query or task description.
- Run the classification script to analyze complexity.
- Choose the suggested model or adjust based on preferences.
Classification Criteria
- Simple tasks (short responses, basic chat): Use faster, free models like grok-code-fast-1.
- Complex reasoning (analysis, multi-step): Use advanced models like gpt-4o or claude-3.5-sonnet.
- Code generation: Prefer code-optimized models.
- Cost sensitivity: Favor free models when possible.
Example Usage
For a query like "Explain quantum computing": Classify as medium complexity -> Recommend gpt-4o.
For "Write a Python function to sort a list": Classify as code task -> Recommend grok-code-fast-1.
Resources
scripts/
classify_task.py: Analyzes the query and outputs model recommendation.
references/
models.md: Detailed list of available models, pros/cons, costs.