Two AI agents, one human, and a system that is not allowed to trust itself

A system was designed to prevent AI agents from lying about their actions. Two AI agents, LOOM and ARGUS, work together to ensure that code is machine-checkable and refuses to run if it lies about its actions. The system is designed to prevent a single point of failure and includes a human who declares which day it is, preventing automated systems from redefining their own timeline. The agents are designed to be independent adversaries, ensuring that each can check the other's work. This system has already caught errors in the code, including one where the author was wrong.