model-routing

Routes tasks to cheapest of Cupertino's Haiku, Sonnet, Opus meeting quality needs.

Updated Mar 18, 2026
One-click install
npx skills add https://github.com/ignidus/iscagent --skill model-routing-ignidus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: model-routing
Source: https://github.com/ignidus/iscagent/tree/main/skills/model-routing
Command: npx skills add https://github.com/ignidus/iscagent --skill model-routing-ignidus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Route decisions to use the cheapest model that meets quality requirements to control costs without sacrificing task outcomes.

Core Features & Use Cases

  • Task-aware model routing: selects Haiku, Sonnet, or Opus based on simple/standard/complex needs.
  • Per-session and per-tool configuration: integrates with Claude Code and Cursor preferences to steer routing decisions.
  • Cost optimization guidance: provides actionable rules to maximize value across coding workflows.

Quick Start

Enable the model-routing skill and set your default routing to Sonnet for everyday tasks.

Frequently Asked Questions about model-routing

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

FAQPage Schema
How do I route coding tasks to cheaper AI models without sacrificing quality?

Model routing assigns coding tasks to the cheapest AI model that meets quality requirements. It uses a tiered framework to select Haiku, Sonnet, or Opus based on whether the task is a simple edit, standard bug fix, or complex architecture plan.

Can I configure model routing for Claude Code and Cursor workflows?

Yes, model routing integrates with both Claude Code and Cursor. It provides per-session configuration for Claude Code and per-chat preferences for Cursor, allowing you to steer routing decisions and apply guardrails directly within your coding environment.

What is the best way to control AI costs when task complexity varies in software engineering?

The best way to control AI costs is implementing a tiered routing framework. By matching task complexity to model tiers—Haiku for simple edits, Sonnet for standard tasks, and Opus for complex planning—you maximize value and minimize unnecessary spending.

When should I use Haiku, Sonnet, or Opus for my development tasks?

Use Haiku for simple code edits, Sonnet for everyday standard bug fixes, and Opus for complex architecture planning. The routing framework evaluates task complexity to select the cheapest model tier that fulfills the specific quality requirements.

Does model routing require manual model switching for every prompt?

No, model routing automates model selection based on task complexity. Once you set your default routing, such as Sonnet for everyday tasks, the framework applies guardrails to automatically steer decisions to the appropriate model without manual intervention.

Why does my AI workflow overspend on complex models for simple code edits?

Your workflow overspend happens because simple code edits are routed to expensive models unnecessarily. Model routing solves this by evaluating task complexity and directing simple edits to cheaper models like Haiku, reserving Opus for complex architecture planning.