FJP-CONF: an open conformance standard for auditable agent decisions

i don't think many have considered this aspect of ai agent building...yet. the cost of being wrong and unauditable