One-click install
npx skills add https://github.com/cogni-work/cogni-works --skill claims-cogni-work
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claims
Source: https://github.com/cogni-work/cogni-works/tree/main/cogni-claims/skills/claims
Command: npx skills add https://github.com/cogni-work/cogni-works --skill claims-cogni-work

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill systematically verifies claims made in content against their cited sources, catching errors, misquotations, and unsupported conclusions before they are published.

Core Features & Use Cases

  • Automated Verification: Compares statements against provided URLs to detect discrepancies.
  • Deviation Detection: Identifies issues like misquotations, unsupported conclusions, and data staleness.
  • Use Case: After an AI generates a report with cited sources, use this Skill to automatically check if the AI accurately represented the information in those sources, ensuring factual integrity.

Quick Start

Verify all pending claims by running the verify command.

Frequently Asked Questions about claims

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

FAQPage Schema
How do I fact-check LLM-generated content against cited sources?

To fact-check LLM-generated content, this Skill compares statements against provided URLs to detect discrepancies. It automates verifying sourced claims by comparing AI statements with source excerpts to prevent misrepresentations.

What is source verification for AI claims and how does it work?

Source verification for AI claims is the process of comparing statements against cited URLs. This Skill manages the lifecycle of sourced claims, detecting deviations like misquotations and unsupported conclusions by checking statements against source excerpts.

Can I detect unsupported conclusions in AI reports automatically?

Yes, you can detect unsupported conclusions automatically. The Skill identifies deviations including misquotations, unsupported conclusions, and data staleness by comparing claims with cited URL excerpts to ensure content integrity.

What's the best way to verify sourced claims before publishing?

The best way to verify sourced claims is running the verify command. This systematically checks pending claims, comparing statements against provided URLs to catch errors, misquotations, and unsupported conclusions before publication.

Do I need to provide URLs to check claims for factual integrity?

Yes, you need to provide URLs to check claims for factual integrity. The Skill verifies submitted claims by comparing them against cited source URLs to detect any deviations in the AI-generated content.

Why does AI content need claim verification against cited sources?

AI content needs claim verification because AI can misrepresent information. This Skill ensures factual integrity by catching errors, misquotations, and unsupported conclusions through automated verification against cited URLs.