Cover Image

How Safe Are AI-Generated Patches? A Large-scale Study on Security Risks in LLM and Agentic Automated Program Repair on SWE-bench

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Sajadi, Amirali, Damevski, Kostadin, Chatterjee, Preetha
Format:	Preprint
Published:	2025
Subjects:	Cryptography and Security Machine Learning Software Engineering
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Towards Personalizing Secure Programming Education with LLM-Injected Vulnerabilities
by: Frazier, Matthew, et al.
Published: (2026)

AXE: An Agentic eXploit Engine for Confirming Zero-Day Vulnerability Reports
by: Sajadi, Amirali, et al.
Published: (2026)

Psycholinguistic Analyses in Software Engineering Text: A Systematic Literature Review
by: Sajadi, Amirali, et al.
Published: (2025)

Do LLMs Consider Security? An Empirical Study on Responses to Programming Questions
by: Sajadi, Amirali, et al.
Published: (2025)

Automated Repair of TEE Partitioning Issues via DSL-Guided and LLM-Assisted Patching
by: Ma, Chengyan, et al.
Published: (2026)

Accelerating Automatic Program Repair with Dual Retrieval-Augmented Fine-Tuning and Patch Generation on Large Language Models
by: Guo, Hanyang, et al.
Published: (2025)

Secure Coding with AI -- From Detection to Repair
by: Belozerov, Vladislav, et al.
Published: (2025)

LLM4CVE: Enabling Iterative Automated Vulnerability Repair with Large Language Models
by: Fakih, Mohamad, et al.
Published: (2025)

A Systematic Study of LLM-Based Architectures for Automated Patching
by: Xu, Qingxiao, et al.
Published: (2026)

Automated Repair of OpenID Connect Programs (Extended Version)
by: Rahat, Tamjid Al, et al.
Published: (2025)

Toward Automated Security Risk Detection in Large Software Using Call Graph Analysis
by: Pecka, Nicholas, et al.
Published: (2025)

RealSec-bench: A Benchmark for Evaluating Secure Code Generation in Real-World Repositories
by: Wang, Yanlin, et al.
Published: (2026)

Patch-to-PoC: A Systematic Study of Agentic LLM Systems for Linux Kernel N-Day Reproduction
by: Pu, Juefei, et al.
Published: (2026)

How Agentic AI Coding Assistants Become the Attacker's Shell
by: Liu, Yue, et al.
Published: (2026)

From LLMs to Agents: A Comparative Evaluation of LLMs and LLM-based Agents in Security Patch Detection
by: Han, Junxiao, et al.
Published: (2025)

Poisoning Programs by Un-Repairing Code: Security Concerns of AI-generated Code
by: Improta, Cristina
Published: (2024)

How Secure is Secure Code Generation? Adversarial Prompts Put LLM Defenses to the Test
by: Tessa, Melissa, et al.
Published: (2026)

APPATCH: Automated Adaptive Prompting Large Language Models for Real-World Software Vulnerability Patching
by: Nong, Yu, et al.
Published: (2024)

Security in the Age of AI Teammates: An Empirical Study of Agentic Pull Requests on GitHub
by: Siddiq, Mohammed Latif, et al.
Published: (2026)

SecureFixAgent: A Hybrid LLM Agent for Automated Python Static Vulnerability Repair
by: Gajjar, Jugal, et al.
Published: (2025)

Data Augmentation for Improving Emotion Recognition in Software Engineering Communication
by: Imran, Mia Mohammad, et al.
Published: (2022)

VulKey: Automated Vulnerability Repair Guided by Domain-Specific Repair Patterns
by: Li, Jia, et al.
Published: (2026)

Root-Cause-Driven Automated Vulnerability Repair
by: Wang, Hulin, et al.
Published: (2026)

How to Compare the Security of Code Written by Humans to LLM-generated Code
by: Balebako, Rebecca, et al.
Published: (2026)

When "Correct" Is Not Safe: Can We Trust Functionally Correct Patches Generated by Code Agents?
by: Peng, Yibo, et al.
Published: (2025)

When AI Takes the Wheel: Security Analysis of Framework-Constrained Program Generation
by: Liu, Yue, et al.
Published: (2025)

Heimdallr: Characterizing and Detecting LLM-Induced Security Risks in GitHub CI Workflows
by: Ruan, Bonan, et al.
Published: (2026)

Similar but Patched Code Considered Harmful -- The Impact of Similar but Patched Code on Recurring Vulnerability Detection and How to Remove Them
by: Tan, Zixuan, et al.
Published: (2024)

Towards Demystifying and Repairing LLM-in-the-Loop Vulnerabilities
by: Ma, Yujie, et al.
Published: (2026)

StriderSPD: Structure-Guided Joint Representation Learning for Binary Security Patch Detection
by: Li, Qingyuan, et al.
Published: (2026)

AgenticCyOps: Securing Multi-Agentic AI Integration in Enterprise Cyber Operations
by: Mitra, Shaswata, et al.
Published: (2026)

Evaluating Tool Cloning in Agentic-AI Ecosystems
by: Kim, Taein, et al.
Published: (2026)

LLM Security Guard for Code
by: Kavian, Arya, et al.
Published: (2024)

SafeTrans: LLM-assisted Transpilation from C to Rust
by: Farrukh, Muhammad, et al.
Published: (2025)

Threat Modelling and Risk Analysis for Large Language Model (LLM)-Powered Applications
by: Tete, Stephen Burabari
Published: (2024)

Fixing 7,400 Bugs for 1$: Cheap Crash-Site Program Repair
by: Zheng, Han, et al.
Published: (2025)

PatchFuzz: Patch Fuzzing for JavaScript Engines
by: Wang, Junjie, et al.
Published: (2025)

VulnRepairEval: An Exploit-Based Evaluation Framework for Assessing Large Language Model Vulnerability Repair Capabilities
by: Wang, Weizhe, et al.
Published: (2025)

AIM: Automated Input Set Minimization for Metamorphic Security Testing
by: Chaleshtari, Nazanin Bayati, et al.
Published: (2024)

Empirical Study of Code Large Language Models for Binary Security Patch Detection
by: Li, Qingyuan, et al.
Published: (2025)