# VulnRepairEval: An Exploit-Based Evaluation Framework for Assessing Large Language Model Vulnerability Repair Capabilities

arXiv: https://arxiv.org/abs/2509.03331 · HTML: https://mlsanity.com/paper/2509.03331/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Sep 2025.

## Built on by (most cited first, 12 of 12)

- [SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?](https://mlsanity.com/paper/2605.26548/) (2026)
- [ZeroDayBench: Evaluating LLM Agents on Unseen Zero-Day Vulnerabilities for Cyberdefense](https://mlsanity.com/paper/2603.02297/) (2026)
- [EvoRepair: Enhancing Vulnerability Repair Agents Through Experience-Based Self-Evolution](https://mlsanity.com/paper/2605.30105/) (2026)
- [CIPHER: Cryptographic Insecurity Profiling via Hybrid Evaluation of Responses](https://mlsanity.com/paper/2602.01438/) (2026)
- [AI In Cybersecurity Education -- Scalable Agentic CTF Design Principles and Educational Outcomes](https://mlsanity.com/paper/2603.21551/) (2026)
- [When Passing Tests Hides Vulnerabilities: An Empirical Study of Silent Failures in Agentic Systems](https://mlsanity.com/paper/2609.10548/) (2026)
- [Evaluating LLMs for One-Shot Patching of Real and Artificial Vulnerabilities](https://mlsanity.com/paper/2511.23408/) (2025)
- [SoK: AI Secure Code Generation: Progress, Pitfalls, and Paths Forward](https://mlsanity.com/paper/2606.25195/) (2026)
- [Helpful or Harmful? Evaluating LLM-Assisted Vulnerability Patching via a Human Study](https://mlsanity.com/paper/2606.25973/) (2026)
- [The History Is the Detector: Executing CVE Patch History, End-to-End](https://mlsanity.com/paper/2609.05335/) (2026)
- [Beyond Static Guarantees: Measuring the Static-Pass Dynamic-Fail Gap in Security-Sensitive and LLM-Generated Python Code](https://mlsanity.com/paper/2609.10762/) (2026)
- [P2Patch: Exploit-Guided Vulnerability Repair](https://mlsanity.com/paper/2610.20245/) (2026)

## Abstract (arXiv)

The adoption of Large Language Models (LLMs) for automated software vulnerability patching has shown promising outcomes on carefully curated evaluation sets. Nevertheless, existing datasets predominantly rely on superficial validation methods rather than exploit-based verification, leading to overestimated performance in security-sensitive applications. This paper introduces VulnRepairEval, an evaluation framework anchored in functional Proof-of-Concept (PoC) exploits. Our framework delivers a comprehensive, containerized evaluation pipeline that enables reproducible differential assessment, where repair success requires the original exploit to fail execution against the modified code. The benchmark construction involved extensive data curation: we processed over 400 CVEs and approximately 2,500 potential sources to extract a collection of authentic vulnerability instances (23 Python CVEs) amenable to automated testing with working PoCs. Through VulnRepairEval, we conduct a comprehensive evaluation of 12 popular LLMs and observe a significant performance deficit: even the top-performing model successfully addresses merely 5/23 instances (about 21.7%), exposing critical weaknesses in security-focused applications. Our failure analysis reveals that most unsuccessful attempts stem from imprecise vulnerability identification and patches containing syntactic or semantic errors. Enhanced prompting strategies and multi-agent approaches yield minimal improvements, with overall effectiveness remaining largely unaffected. This work contributes a stringent, practical evaluation framework for LLM-driven vulnerability remediation and underscores the necessity for assessment protocols that authentically reflect real-world exploitation scenarios.