PartitionGPT is a novel approach that leverages large language models to automatically partition smart contracts, protecting sensitive information from manipulation attacks.

• First-of-its-kind solution combining static analysis with LLMs' in-context learning capabilities
• Prevents information leakage by intelligently separating public from private contract components
• Addresses fundamental vulnerabilities stemming from data confidentiality issues rather than implementation bugs
• Enhances blockchain security by protecting contracts from manipulation without sacrificing functionality

This research represents a significant advancement in smart contract security, offering practical protection against manipulation attacks that exploit sensitive information exposure in blockchain applications.

Towards Secure Program Partitioning for Smart Contracts with LLM's In-Context Learning