unbiased-test

Compare two data sources and report all discrepancies.

8|1|Updated Aug 5, 2025

Search Tags:#verification #auditing #data-migration #diff-check #data-discrepancies #fresh-instance #unbiased

Ready to run

npx skills add https://github.com/ArtyMcLabin/ClaudeCode-global-config --skill unbiased-test

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: unbiased-test
Source: https://github.com/ArtyMcLabin/ClaudeCode-global-config/tree/main/skills/unbiased-test
Command: npx skills add https://github.com/ArtyMcLabin/ClaudeCode-global-config --skill unbiased-test

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps teams verify that two data sources are consistent by performing an unbiased, fresh-instance audit that reveals all discrepancies without assuming which source is correct.

Core Features & Use Cases

Fresh Claude Code session for each run to avoid context leakage.
Adversarial framing to uncover hidden mismatches between datasets, migrations, or transformed files.
Blind comparison and comprehensive discrepancy reporting to support data validation and quality assurance.

Quick Start

Run the unbiased-test skill to compare two datasets or data sources and report all differences.

Frequently Asked Questions about unbiased-test

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

FAQPage Schema

How do I verify data consistency after a data migration?▼

Verify data consistency after a migration by running an unbiased audit in a fresh Claude session that performs blind comparisons to reveal all data discrepancies without assuming which source is correct.

Can I check for data discrepancies between two transformed files?▼

Yes, you can check for data discrepancies between two transformed files by applying a neutral prompt and iterating re-verification until zero mismatches remain. This supports data validation and quality assurance.

Does verifying data discrepancies require a fresh session to avoid bias?▼

Yes, verifying data discrepancies requires a fresh Claude session for each run to avoid context leakage and ensure a truly unbiased, adversarial evaluation of the two data sources.

What is the best way to audit data sources without assuming which is correct?▼

The best way to audit data sources without assuming correctness is performing a blind comparison using an adversarial framing, which uncovers hidden mismatches between datasets, migrations, or transformed files.

When do I need iterative re-verification for data migration testing?▼

You need iterative re-verification for data migration testing when you must ensure zero discrepancies between two data sources, requiring repeated fresh evaluations until all differences are resolved.