test-validation

This project implements an AI agent that verifies if automated Hercules test runs were executed as intended by comparing planning logs, video evidence, and final outputs. It uses open-source LLMs and computer vision tools to flag deviations, providing detailed reports with technical insights.