This reminds me of Automated Reasoning where the model maintains a form of lean code which gets run to verify assumptions constantly as the model builds a knowl

This reminds me of Automated Reasoning where the model maintains a form of lean code which gets run to verify assumptions constantly as the model builds a knowledge base