dingo-data-quality

What problem does it solve?

This Skill helps you evaluate the quality, completeness, safety, and factual reliability of your datasets and generated content so you can detect issues before deployment or publication.

Core Features & Use Cases

• Run deterministic data quality checks (e.g., null/format/content rules) to catch formatting errors, missing fields, and basic quality problems quickly.
• Run LLM-based semantic and RAG quality metrics (e.g., text quality, repeat detection, faithfulness, context precision/recall, security prohibition) to evaluate meaning, retrieval grounding, and response quality.
• Run autonomous fact-checking on articles using ArticleFactChecker, producing an accuracy score and structured false-claim/evidence reports for review.

Quick Start

Create a config.json for your JSONL dataset and run Dingo with the command dingo eval --input config.json.