skills.rest
gozonerd
gozonerdCOMMUNITY
Explore all skills by this author →
GitHub

Factual Claim Ground Truth Verification

Extract and verify factual claims in deliverables against authoritative evidence sources.

Updated Apr 11, 2026
Search Tags:#quality assurance#validation#yaml#evidence#ground truth#claim verification#fact checking
One-click install
npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill factual-claim-ground-truth-verification-gozonerd
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Factual Claim Ground Truth Verification
Source: https://github.com/gozonerd/shadow-ai-assessment/tree/main/.claude/skills/dmis/task_type/TT-11_Factual_Claim_Ground_Truth_Verification
Command: npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill factual-claim-ground-truth-verification-gozonerd

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Deliverables often contain factual assertions about statistics, organizations, regulations, and events that may be inaccurate or unverifiable. This Skill systematically extracts every factual claim and verifies each one against authoritative sources, producing a structured YAML verification report with confidence ratings. ## Core Features & Use Cases - Claim Extraction and Categorization: Identifies all factual assertions in a deliverable and organizes them by category such as statistics, organizational facts, and regulatory facts. - Evidence-Based Verification: Checks each claim against an evidence library and ground truth standards, enforcing 9 critical accuracy rules (ACC-001 through ACC-009). - Structured YAML Reporting: Produces a verification report conforming to the TT-11 output schema, with claim inventory, verification results, discrepancy analysis, and confidence ratings, validated by an included Python script. - Use Case: After generating a research report about an organization's technology deployments, run this Skill to confirm every statistic and organizational fact is supported by cited evidence before publication. ## Quick Start Verify all factual claims in the attached deliverable against the evidence library and produce a TT-11 verification report in YAML format.

Frequently Asked Questions about Factual Claim Ground Truth Verification

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

FAQPage Schema
How do I verify factual claims in a document against source material?▼

Extract every factual assertion from the deliverable, categorize claims by type, then check each against authoritative sources in the evidence library. Record the verification result, evidence source, and a confidence rating for each claim in a structured YAML report.

What output format does factual claim verification produce?▼

The verification report is produced in YAML conforming to the TT-11 output schema. It includes claim inventory, verification results, evidence assessment, discrepancy analysis, and confidence ratings sections.

How is the TT-11 verification report validated?▼

Run the included validate_output.py script with the report file path. It checks YAML parsing, task_type, version, and confirms all nine accuracy rules ACC-001 through ACC-009 appear in the accuracy_compliance section.

What are the ACC accuracy rules in claim verification?▼

They are nine critical accuracy rules (ACC-001 through ACC-009) defining ground truth constraints, such as correct terminology, convergence thresholds, and organizational facts. Every verification report must reference all nine rules in its compliance statement.

What happens when a factual claim cannot be verified?▼

Claims that conflict with evidence or lack accessible sources are flagged in the discrepancy analysis section. Confidence ratings reflect evidence quality, so unverifiable claims receive low confidence rather than being marked as verified.

Related Skills

Explore More Content & Communication Skills→
r-irber-irbe
2

nightly-testing

Redirect nightly testing guidance to the upstream reference document.

Community
Basic
Phife726Phife726

fact-checker

Verify factual claims against credible evidence with confidence ratings.

Community
Advanced
openJiuwen-aiopenJiuwen-ai
64

evidence-extractor

Extracts structured, source-grounded evidence items from academic source packages for downstream synthesis.

Official
Advanced
inference-shinference-sh
688

ai-rag-pipeline

Build RAG pipelines integrating web search tools with large language models.

Official
Intermediate