"Code and data for the manuscript entitled \u201cThe Self-Aware Causal AI\u201d"

"Contemporary artificial intelligence excels at pattern recognition but fundamentally fails to distinguish causation from correlation\u2014a crisis that undermines its reliability in science, medicine, and autonomous systems. Here we introduce a first-principles, parameter-free framework that bridges measure-theoretic axioms to executable code, enabling rigorous causal reasoning from purely observational data. We prove causal asymmetry is a topological invariant, yielding a single mathematical construct that simultaneously self-discovers a theoretical discernibility boundary and executes parameter-free directional inference. Evaluated on 108 real-world causal pairs (214,680 samples), our framework achieves 82.86% accuracy within the high-discernibility regime, while providing transparent, mathematical criteria for warranting causal claims. This work establishes self-aware causal AI as an axiomatic engineering discipline, delivering verifiable causal reasoning derived from first principles."