Skill UI

This skill provides comprehensive tools to detect data poisoning and model backdooring attacks throughout the machine learning lifecycle. It checks data provenance, detects label-flip anomalies (using Cleanlab), and analyzes model activations and spectra (using ART) to identify malicious triggers or compromised weights, ensuring ML supply chain integrity.