smart-routing

Route user requests to appropriate agents based on intent and context.

90|18|Updated Nov 27, 2025
One-click install
npx skills add https://github.com/wasintoh/toh-framework --skill smart-routing
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: smart-routing
Source: https://github.com/wasintoh/toh-framework/tree/main/src/skills/smart-routing
Command: npx skills add https://github.com/wasintoh/toh-framework --skill smart-routing

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI often struggles with ambiguous user requests or complex workflows, leading to confusion or incorrect actions. This Skill acts as an intelligent router, ensuring every request is directed to the most appropriate agent or action, streamlining your interaction with the AI.

Core Features & Use Cases

  • Intent Classification Matrix: Analyzes keywords, context, and memory to accurately determine user intent (e.g., create UI, fix bug, connect backend).
  • Confidence Scoring Algorithm: Assesses the clarity of the request, deciding whether to execute directly, plan first, or ask for clarification, ensuring efficient task handling.
  • IDE-Aware Execution: Adapts multi-agent strategy based on the detected IDE (e.g., parallel execution in Claude Code, sequential in others) for optimized performance.
  • Use Case: You simply say 'fix it', and this skill, leveraging project memory, understands you mean 'fix the recently introduced bug on the dashboard' and routes the request to the Fix Agent, without needing further clarification, saving you time.

Quick Start

Use the smart-routing skill to handle the request 'build e-commerce'.

Frequently Asked Questions about smart-routing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I route user requests to the right agent based on intent?

Intent-based routing analyzes keywords, context, and memory to classify what users actually need—like distinguishing 'fix it' as 'fix the dashboard bug'—then automatically directs the request to the appropriate agent without requiring clarification.

Can I use request routing with multiple agents in a workflow?

Yes. Multi-agent orchestration with confidence-based decision making adapts execution strategy by IDE type: parallel in Claude Code for speed, sequential elsewhere for compatibility, ensuring efficient task handling across agent networks.

How does confidence scoring help with ambiguous requests?

Confidence scoring assesses request clarity to decide whether to execute directly, plan first, or ask for clarification, preventing incorrect actions on unclear user input and streamlining decision logic in complex workflows.

What's the difference between direct execution and planning-first routing?

Routing compares request confidence against thresholds: high-confidence requests execute immediately to the right agent, while lower-confidence requests route to planning mode first, balancing speed with accuracy in task handling.

Does request routing work with memory-driven context?

Yes. Smart routing enforces memory-first context building from the memory store, extracting project history and prior state to disambiguate requests and make accurate intent classifications without repeated clarification.

Can I adapt routing strategy based on my IDE environment?

Yes. IDE detection automatically adjusts multi-agent strategy—optimizing for parallel execution in Claude Code or sequential execution in other IDEs—ensuring routing decisions match your development environment's capabilities.