Hypothesis-Disciplined Multi-Agent Automated Formalization of Asymptotic Statistical Theory

A new formalization pipeline for asymptotic statistical theory has been developed using Lean 4, involving a multi-agent system that coordinates seven specialized roles for proof planning and construction. The approach emphasizes a hypothesis-disciplined audit process, ensuring that each theorem's hypotheses are rigorously justified or rejected, resulting in an axiom-clean and source-faithful Lean development. This systematic formalization, which includes Lean-checked proofs of core parametric and semi-parametric theorems, addresses significant gaps in existing formalization infrastructure and is accessible via GitHub.