“ProofSketcher combines large language models with lightweight proof checkers to catch mathematical and logical errors that LLMs typically conceal. This hybrid approach addresses a critical reliability gap in AI reasoning systems, enabling more trustworthy applications in domains requiring rigorous logical correctness.”