Newest papers

How Effectively Do Security Skills Guide Coding Agents in Vulnerability Detection in Linux Kernel?

Haoran Xue, Nikta Akbarpour, Gias Uddin, Song Wang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.25705 · cited by 0 works (OpenAlex) · open-access PDF

French

Les agents de codage de vulnérabilité potentielle (LLM), tels que Claude Code, sont de plus en plus utilisés pour l'audit de sécurité, avec des compétences spécifiques aux tâches fournissant des procédures d'analyse de vulnérabilité réutilisables. Cependant, leur efficacité dans la détection de vulnérabilités réelles et les raisons de leurs échecs demeurent incertaines. Nous menons la première étude empirique à grande échelle de 37 compétences de sécurité en utilisant Claude Code avec quatre LLM sous-jacentes sur 100 vulnérabilités du noyau Linux réel sur dix classes, soit un total de 14 800 cours. Nous évaluons la détection de vulnérabilité à des niveaux de fonction, déclaration et classe de vulnérabil

Machine translation (Meta M2M100); the English original follows.

Abstract

Large language model (LLM) coding agents, such as Claude Code, are increasingly used for security auditing, with task-specific skills providing reusable vulnerability-analysis procedures. However, their effectiveness in detecting real vulnerabilities and the reasons for their failures remain unclear. We conduct the first large-scale empirical study of 37 security skills using Claude Code with four underlying LLMs on 100 real Linux kernel vulnerabilities across ten classes, totaling 14800 runs. We evaluate vulnerability detection at the function, statement, and vulnerability-class levels. Results show that 24.4 percent of runs locate the correct function, 18.3 percent identify both the function and responsible statement, and 15.7 percent correctly classify the vulnerability. Our failure analysis reveals limitations in evidence acquisition, state-relation reasoning, code interpretation, candidate verification, investigation coverage, and reporting. Based on these findings, we propose SKILL-SIF, a self-improving framework that iteratively evolves security skills by learning from missed vulnerabilities and failure patterns. We evolve five top-performing skills using the original benchmark and evaluate them on 100 previously unseen vulnerabilities. SKILL-SIF improves fully correct vulnerability detection by up to eight times (from 2 percent to 16 percent) on unseen vulnerabilities, while aggregate accuracy across four LLMs increases by 1.85 times (from 6.05 percent to 11.20 percent). These findings highlight the limitations of existing security skills and demonstrate the potential of failure-driven skill evolution to improve vulnerability detection in coding agents.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Vulnerability (computing), Coding (social sciences), Benchmark (surveying), Computer security, Secure coding

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange