realitycheck-extract

Extract discrete claims from a source into a structured YAML list with metadata fields.

38|Updated Jan 20, 2026

Search Tags:#evidence #sources #claims #knowledge-base #extract #rapid-analysis

Ready to run

npx skills add https://github.com/lhl/realitycheck --skill realitycheck-extract

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: realitycheck-extract
Source: https://github.com/lhl/realitycheck/tree/main/integrations/amp/skills/realitycheck-extract
Command: npx skills add https://github.com/lhl/realitycheck --skill realitycheck-extract

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables rapid extraction of claims from a source without performing full 3-stage analysis, reducing manual triage time.

Core Features & Use Cases

Targeted extraction of discrete claims from a document or passage.
Evidence tagging to link claims to supporting sources.
Lightweight workflow for quick triage and cross-source reference.

Quick Start

Provide the source text or URL to extract claims from; the tool will return a claims list in YAML with fields id, text, type, domain, evidence_level, credence, and source_ids.

Frequently Asked Questions about realitycheck-extract

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

FAQPage Schema

How do I extract claims from a document for rapid analysis?▼

To extract claims from a document, provide the source text or URL. The tool returns a structured YAML list of claims with fields like id, text, type, domain, evidence_level, credence, and source_ids for quick triage.

What is the best way to harvest supporting evidence across multiple sources?▼

Harvesting supporting evidence across sources is done by running targeted extractions on each source. This lightweight workflow outputs a structured claims list with evidence tags and source_ids, enabling cross-source reference without full analysis.

Can I tag evidence to link claims back to their original sources?▼

Yes, you can tag evidence to link claims to supporting sources. The extraction output includes a source_ids field within the YAML claims list, ensuring every harvested claim maintains its link to the original document.

How do I quickly triage claims without performing a full 3-stage analysis?▼

To triage claims without a full 3-stage analysis, use this rapid extraction skill. It bypasses deep processing and directly outputs a structured YAML list of discrete claims with evidence levels and credence, reducing manual triage time.

Are there limitations to using rapid claim extraction for knowledge-base building?▼

The main limitation of rapid claim extraction is that it skips full 3-stage analysis, meaning it is designed for quick triage and gathering evidence rather than providing deep, comprehensive validation of the claims within your knowledge-base.