verification-retry-loop

What problem does it solve?

This Skill ensures the quality of data extracted from LLMs by automatically verifying it and retrying if necessary, with a maximum of one retry, to maintain accuracy.

Core Features & Use Cases

• Structured Data Extraction: Supports structured Pydantic extraction from LLMs.
• Deterministic Post-Checks: Runs deterministic checks on the output.
• Bounded Retries: Automatically retries extraction with feedback if issues are detected, up to one time.
• Use Case: Ideal for extracting information from user documents and ensuring the data quality before final processing.

Quick Start

Implement the extract_with_verification function with your Pydantic model and verification function.