Use a Vision Transformer in code repositories to perform detailed image analysis using advanced search and clustering techniques.
Scan the product manifest — ontology types, functions, actions, policies, and dataset shells.