Comment by tiffanyh

> Then AIP reads the agent's reasoning between every action and compares it to that contract.

How would this work? Is one LLM used to “read” (and verify) another LLMs reasoning?