Towards Repository-Level Program Verification with Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhong, Si Cheng, Si, Xujie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908566658482176
author Zhong, Si Cheng
Si, Xujie
author_facet Zhong, Si Cheng
Si, Xujie
contents Recent advancements in large language models (LLMs) suggest great promises in code and proof generations. However, scaling automated formal verification to real-world projects requires resolving cross-module dependencies and global contexts, which are crucial challenges overlooked by existing LLM-based methods with a special focus on targeting isolated, function-level verification tasks. To systematically explore and address the significant challenges of verifying entire software repositories, we introduce RVBench, the first verification benchmark explicitly designed for repository-level evaluation, constructed from four diverse and complex open-source Verus projects. We further introduce RagVerus, an extensible framework that synergizes retrieval-augmented generation with context-aware prompting to automate proof synthesis for multi-module repositories. RagVerus triples proof pass rates on existing benchmarks under constrained model inference budgets, and achieves a 27% relative improvement on the more challenging RVBench benchmark, demonstrating a scalable and sample-efficient verification solution.
format Preprint
id arxiv_https___arxiv_org_abs_2509_25197
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Repository-Level Program Verification with Large Language Models
Zhong, Si Cheng
Si, Xujie
Software Engineering
Artificial Intelligence
Programming Languages
Recent advancements in large language models (LLMs) suggest great promises in code and proof generations. However, scaling automated formal verification to real-world projects requires resolving cross-module dependencies and global contexts, which are crucial challenges overlooked by existing LLM-based methods with a special focus on targeting isolated, function-level verification tasks. To systematically explore and address the significant challenges of verifying entire software repositories, we introduce RVBench, the first verification benchmark explicitly designed for repository-level evaluation, constructed from four diverse and complex open-source Verus projects. We further introduce RagVerus, an extensible framework that synergizes retrieval-augmented generation with context-aware prompting to automate proof synthesis for multi-module repositories. RagVerus triples proof pass rates on existing benchmarks under constrained model inference budgets, and achieves a 27% relative improvement on the more challenging RVBench benchmark, demonstrating a scalable and sample-efficient verification solution.
title Towards Repository-Level Program Verification with Large Language Models
topic Software Engineering
Artificial Intelligence
Programming Languages
url https://arxiv.org/abs/2509.25197