golden-dataset-validation

What problem does it solve?

This skill validates golden datasets by enforcing schema rules, detecting duplicates, and ensuring referential integrity, thereby preventing data quality issues in AI workflows.

Core Features & Use Cases

• Schema validation for documents and queries to guarantee consistent structure.
• Duplicate detection via semantic similarity and URL checks to maintain dataset uniqueness.
• Coverage analysis and referential checks to reveal gaps and preserve relationships between documents and queries.
• Use Case: Pre-commit validation before adding new items to the golden dataset, and full-dataset audits.

Quick Start

Run full validation with the provided scripts to inspect new and existing dataset entries. For example, execute the pre-commit hook or the validate-all workflow to surface errors and warnings, then address issues before committing.