smart-router
Expertise-aware model router with semantic domain scoring, context-overflow protection, and security redaction. Automatically selects the optimal AI model using weighted expertise scoring (Feb 2026 benchmarks). Supports Claude, GPT, Gemini, Grok with automatic fallback chains, HITL gates, and cost optimization.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/c0nspic0us7urk3r/smart-routerWhat This Skill Does
The smart-router skill is an advanced orchestrator designed to optimize AI performance, cost, and latency within the OpenClaw ecosystem. Rather than relying on a single static model, smart-router evaluates every user query against a three-tier decision matrix. In the first tier, it analyzes intent to categorize the task into domains like CODE, ANALYSIS, CREATIVE, REALTIME, or GENERAL. The second tier assesses the complexity of the task, distinguishing between simple, one-step queries and complex, multi-step reasoning requirements. Finally, the third tier applies overrides for specific constraints, such as massive context windows or specialized real-time data requirements.
By dynamically routing prompts to the most capable yet cost-effective model (such as Haiku for simple tasks or Opus/GPT-5 for complex reasoning), it ensures the user receives high-quality results without unnecessary latency or over-spending on expensive API tokens. It handles model fallbacks automatically, maintaining seamless continuity for the end user.
Installation
Install the smart-router skill directly via the command line within your OpenClaw environment:
clawhub install openclaw/skills/skills/c0nspic0us7urk3r/smart-router
Use Cases
- Software Engineering: Developers can offload debugging, code refactoring, and documentation generation, knowing the router will select a model optimized for programming logic.
- Large Data Analysis: Users uploading dense documentation or codebase archives benefit from the automatic Gemini Pro fallback when context exceeds 100K tokens.
- Real-time Research: Users requiring the latest breaking news or social sentiment analysis on markets will have their queries automatically routed to Grok, bypassing general-purpose models.
- Cost Efficiency: Enterprises can maintain high quality while significantly reducing token expenditure by utilizing smaller models for simple greetings or factual lookups.
Example Prompts
- "Refactor this Python script to use asynchronous I/O and fix the memory leak in the handle_request function. [show routing]"
- "Summarize the attached 200,000-word industry report and list the top five growth projections for 2026."
- "What is the current stock sentiment on X for the tech sector following today's market opening?"
Tips & Limitations
- Transparency: Use the tag
[show routing]in any message to receive a debug log explaining which model was chosen and why. This is excellent for verifying that the router is functioning as expected for specific project needs. - Explicit Control: The router is designed to be invisible, but if you have a preference for a specific model, your explicit override will always take precedence over the internal scoring logic.
- Constraints: Be aware that while the router handles context-overflow protection, queries exceeding 1M tokens are still subject to model-specific limits. Ensure you are aware of your specific OpenClaw subscription tier.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-c0nspic0us7urk3r-smart-router": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: external-api