VeriStruct: AI-assisted Automated Verification of Data-Structure Modules in Verus

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sun, Chuyue, Sun, Yican, Amrollahi, Daneshvar, Zhang, Ethan, Lahiri, Shuvendu, Lu, Shan, Dill, David, Barrett, Clark
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910039473651712
author Sun, Chuyue
Sun, Yican
Amrollahi, Daneshvar
Zhang, Ethan
Lahiri, Shuvendu
Lu, Shan
Dill, David
Barrett, Clark
author_facet Sun, Chuyue
Sun, Yican
Amrollahi, Daneshvar
Zhang, Ethan
Lahiri, Shuvendu
Lu, Shan
Dill, David
Barrett, Clark
contents We introduce VeriStruct, a novel framework that extends AI-assisted automated verification from single functions to more complex data structure modules in Verus. VeriStruct employs a planner module to orchestrate the systematic generation of abstractions, type invariants, specifications, and proof code. To address the challenge that LLMs often misunderstand Verus' annotation syntax and verification-specific semantics, VeriStruct embeds syntax guidance within prompts and includes a repair stage to automatically correct annotation errors. In an evaluation on eleven Rust data structure modules, VeriStruct succeeds on ten of the eleven, successfully verifying 128 out of 129 functions (99.2%) in total. These results represent an important step toward the goal of automatic AI-assisted formal verification.
format Preprint
id arxiv_https___arxiv_org_abs_2510_25015
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VeriStruct: AI-assisted Automated Verification of Data-Structure Modules in Verus
Sun, Chuyue
Sun, Yican
Amrollahi, Daneshvar
Zhang, Ethan
Lahiri, Shuvendu
Lu, Shan
Dill, David
Barrett, Clark
Software Engineering
Artificial Intelligence
We introduce VeriStruct, a novel framework that extends AI-assisted automated verification from single functions to more complex data structure modules in Verus. VeriStruct employs a planner module to orchestrate the systematic generation of abstractions, type invariants, specifications, and proof code. To address the challenge that LLMs often misunderstand Verus' annotation syntax and verification-specific semantics, VeriStruct embeds syntax guidance within prompts and includes a repair stage to automatically correct annotation errors. In an evaluation on eleven Rust data structure modules, VeriStruct succeeds on ten of the eleven, successfully verifying 128 out of 129 functions (99.2%) in total. These results represent an important step toward the goal of automatic AI-assisted formal verification.
title VeriStruct: AI-assisted Automated Verification of Data-Structure Modules in Verus
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2510.25015