LLM-driven Program Partitioning Prevents Smart Contract Data Leakage
LLM-driven program partitioning isolates sensitive contract data, mathematically preventing manipulation attacks and securing decentralized applications.
LLMs Automate Property Generation for Smart Contract Formal Verification
PropertyGPT leverages large language models and retrieval-augmented generation to automate the creation of formal verification properties, dramatically reducing the manual effort required for smart contract security.
Fuzzing Zero-Knowledge Proof Circuits Ensures Implementation Security and Reliability
Introducing fuzzing to ZKP circuits solves the oracle problem for soundness, establishing a scalable, practical security layer for verifiable computation.
LLM-Driven Property Generation Automates Smart Contract Formal Verification and Auditing
PropertyGPT uses retrieval-augmented LLMs and iterative refinement to automatically generate formal verification properties, fundamentally mitigating the critical human-expertise bottleneck in smart contract security.
LLMs Automate Property Generation, Resolving the Smart Contract Verification Bottleneck
A retrieval-augmented LLM framework automatically generates formal properties, drastically improving the scalability and security assurance of smart contracts.
Formal Verification Secures Zero-Knowledge Proof Circuits from Exploitable Flaws
Novel static analysis and verification tools precisely identify and prevent critical vulnerabilities within zero-knowledge proof circuits, fortifying decentralized systems.
LLMs Automate Smart Contract Formal Verification Property Generation
A novel system leverages large language models and retrieval-augmented generation to automate smart contract property creation, enhancing security and accessibility.
