scoring-engine

Compute official_source_ratio and per-question status for AI platform citations.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/opensourceways/geo-workflow --skill scoring-engine-opensourceways
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scoring-engine
Source: https://github.com/opensourceways/geo-workflow/tree/main/.claude/skills/scoring-engine
Command: npx skills add https://github.com/opensourceways/geo-workflow --skill scoring-engine-opensourceways

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Evaluates AI platform responses to determine how well they cite official URLs, enabling objective quality assessment and targeted improvements across communities and questions.

Core Features & Use Cases

  • Per-question, per-platform URL citation checks using exact URL matching and normalization rules.
  • Aggregated scoring with a 75% threshold to classify status (satisfied / not_cited / no_official_content) and a full summary.
  • Generates structured results, including per-platform matched URLs, counts, and a final scoring report that can drive GEO catalog suggestions.

Quick Start

Run the scoring pipeline against a set of responses and questions to generate scoring-results.json for review.

Frequently Asked Questions about scoring-engine

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

FAQPage Schema
How do I evaluate if AI answers cite official URLs across multiple platforms?

To evaluate if AI answers cite official URLs, you can run an automated scoring pipeline that checks per-question, per-platform URL citations. It computes an official_source_ratio and classifies each answer's citation status based on a 75% threshold.

What is official source ratio in AI answer auditing?

Official source ratio in AI answer auditing is a computed metric quantifying how well AI platform responses cite official content. It determines per-question status, classifying answers as satisfied, not_cited, or no_official_content using a 75% threshold.

How do I generate structured scoring reports for AI platform responses?

You generate structured scoring reports for AI platform responses by running the scoring pipeline against a set of questions and answers. This produces a scoring-results.json file containing per-platform matched URLs, counts, and a full summary for downstream reporting.

Does the URL citation check use exact matching or domain-level matching?

The URL citation check uses exact URL matching combined with normalization rules. This ensures precise identification and quantification of how AI platform answers cite official content across questions and platforms.

Can I use the scoring results to drive GEO catalog actions?

You can use the scoring results to drive GEO catalog suggestions. The pipeline generates structured results with per-platform matched URLs, counts, and a final scoring report formatted specifically for automated scoring pipelines and downstream reporting tools.

What dependencies do I need to run the AI answer scoring pipeline?

No external dependencies are required to run the AI answer scoring pipeline. The Skill operates independently using its internal scripts, references, and assets to perform URL citation checks and generate structured scoring reports.