How do I change my model routing profile?

You can switch between quality, balanced, and budget profiles using the command /gsd:set-profile followed by the profile name.

Does this work for projects not using GSD?

Yes, for non-GSD projects, you can manually configure your preferred profile by creating a config.json file in the ~/.config/model-matchmaker/ directory.

Where can I see why a specific model was recommended?

All routing decisions and the logic behind them are recorded in the decision log located at ~/.local/state/model-matchmaker/decisions.log.

What is effort-aware prompt routing?

It is a system that analyzes the complexity of your prompt to suggest the most appropriate Claude model and reasoning level, helping you balance cost, speed, and performance.

Can I bypass a model suggestion for a single turn?

Yes, you can prefix any prompt with a backslash (\) to skip the suggestion and maintain your current model for that specific turn.

Effort-Aware Model Routing

Name: Effort-Aware Model Routing
Author: jacobcxdev

byjacobcxdev

0•

생산성 및 워크플로

Optimizes Claude model selection and reasoning effort based on prompt complexity and user-defined budget profiles.

The effort-config skill for Claude Code introduces intelligent, automated model routing to ensure you are always using the optimal AI model for every task. By classifying prompts into tiers—ranging from trivial mechanical operations like git renames to complex architectural reasoning—it dynamically suggests the best switch between Claude Haiku, Sonnet, and Opus. Users can choose from Quality, Balanced, or Budget profiles to tailor how the skill recommends model changes, ensuring a perfect balance between high-performance reasoning for deep work and cost-efficiency for simple requests.

주요 기능

01Granular reasoning effort level control

020 GitHub stars

03Profile-based routing (Quality, Balanced, Budget)

04Detailed decision logging for performance analysis

05Automated model switching suggestions (Haiku, Sonnet, Opus)

06Intelligent task complexity classification

사용 사례

01Routing simple git commands and trivial file operations to Haiku to minimize costs.

02Streamlining daily development by using Sonnet for standard implementation tasks with automated effort adjustment.

03Ensuring complex security audits or cross-codebase refactoring tasks use Claude Opus with high reasoning effort.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jacobcxdev/model-matchmaker effort-config

For use in Claude.ai and ChatGPT

주요 기능

01Granular reasoning effort level control

020 GitHub stars

03Profile-based routing (Quality, Balanced, Budget)

04Detailed decision logging for performance analysis

05Automated model switching suggestions (Haiku, Sonnet, Opus)

06Intelligent task complexity classification

사용 사례

01Routing simple git commands and trivial file operations to Haiku to minimize costs.

02Streamlining daily development by using Sonnet for standard implementation tasks with automated effort adjustment.

03Ensuring complex security audits or cross-codebase refactoring tasks use Claude Opus with high reasoning effort.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jacobcxdev/model-matchmaker effort-config

For use in Claude.ai and ChatGPT