Can I use this for prompt engineering?

Yes, it is specifically designed to help prompt engineers identify missing nuances and improve the depth of AI responses.

What do the letter grades represent?

Grades range from A+ (top human expert) to F (not useful). Intermediate grades like B+ represent an untrained human with an above-average IQ.

How does the 1-100 score work?

The score represents the proximity to 'perfect' human output, where 100 is equivalent to a world-class expert working on the task full-time for three months.

What criteria does this skill use to rate AI responses?

The skill evaluates AI output against human expertise levels, comparing the response to what a top 1% global expert would produce given the same instructions.

Is this skill part of the Fabric ecosystem?

Yes, it is a specialized implementation of the 'rate_ai_response' pattern originally created by Daniel Miessler for the Fabric project.

AI Response Evaluator

Name: AI Response Evaluator
Author: bdmorin

bybdmorin

•

データサイエンスとML

Evaluates AI-generated content against human expert standards to provide rigorous quality benchmarks and performance scoring.

The AI Response Evaluator skill implements the renowned Fabric 'rate_ai_response' pattern within Claude Code to help developers and prompt engineers objectively measure output quality. It moves beyond simple feedback by comparing AI performance against specific human benchmarks—ranging from average individuals to the world's top 1% experts. By providing a standardized letter grade (A+ through F), a 1-100 numerical score, and detailed justifications, it serves as a critical tool for prompt optimization, model benchmarking, and high-stakes quality assurance.

主な機能

01Detailed qualitative analysis with five-point justification bullets

02Expert-level benchmarking comparing AI output to top-tier human specialists

031 GitHub stars

04Standardized letter grading system for quick quality assessment

05Seamless integration of Daniel Miessler's Fabric evaluation framework

06Granular 1-100 scoring based on instruction adherence and depth

ユースケース

01Conducting quality assurance on automated AI content pipelines

02Benchmarking complex prompt performance during development

03Comparing model outputs to identify the best LLM for specific expert tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add bdmorin/the-no-shop rate-ai-response

For use in Claude.ai and ChatGPT

主な機能

01Detailed qualitative analysis with five-point justification bullets

02Expert-level benchmarking comparing AI output to top-tier human specialists

031 GitHub stars

04Standardized letter grading system for quick quality assessment

05Seamless integration of Daniel Miessler's Fabric evaluation framework

06Granular 1-100 scoring based on instruction adherence and depth

ユースケース

01Conducting quality assurance on automated AI content pipelines

02Benchmarking complex prompt performance during development

03Comparing model outputs to identify the best LLM for specific expert tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add bdmorin/the-no-shop rate-ai-response

For use in Claude.ai and ChatGPT