SentinelCMS: Proactive Vulnerability Detection in CMS Plugins Using Static Taint Analysis and Bidirectional LSTM
Applied Sciences · 01.06.2026 · 16 (11) · 5471
DOI: 10.3390/app16115471 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 01.06.2026. Verfasst von Zhuldyz Tashenova und 6 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Content Management Systems (CMS) have fundamentally transformed the digital landscape, powering a substantial majority of the modern web. However, the ubiquity of platforms such as WordPress, Joomla, and Drupal has made them primary targets for cybercriminals. The central security weakness lies not within the core software but within the complex ecosystem of third-party extensions and themes, which account for the vast majority of reported vulnerabilities. This article presents a comprehensive analysis of the current CMS security landscape, synthesizing empirical data on the prevalence of outdated components and the efficacy of existing scanning tools. Based on a critical review of recent literature and an analysis of common attack vectors, the study identifies the fundamental limitations of traditional, reactive security paradigms that rely on signature-based detection. To address the critical gap in detecting zero-day threats, this research proposes and implements a proactive vulnerability detection system—SentinelCMS—combining static Taint Analysis with a Bidirectional LSTM neural network classifier. The system was validated on an augmented dataset of 600 PHP code samples using 5-fold cross-validation. An experimental proof of concept validates that static analysis of plugin source code can extract semantic features to accurately classify vulnerabilities into specific categories (SQL Injection, Cross-Site Scripting, Remote Code Execution) with an overall accuracy of 93%. The implementation of this system aims to shift the cybersecurity paradigm from incident response to threat prevention, thereby significantly enhancing the resilience of web resources.
Zitieren
APA
Tashenova, Z., Aitmagambetuly, A., Bayegizova, A., Santeyeva, S., Abdugulova, Z., Amanzholova, S., & Kerim, A. (2026). SentinelCMS: Proactive Vulnerability Detection in CMS Plugins Using Static Taint Analysis and Bidirectional LSTM. Applied Sciences, 16(11), 5471. https://doi.org/10.3390/app16115471
Vancouver
Tashenova Z, Aitmagambetuly A, Bayegizova A, Santeyeva S, Abdugulova Z, Amanzholova S, et al. SentinelCMS: Proactive Vulnerability Detection in CMS Plugins Using Static Taint Analysis and Bidirectional LSTM. Applied Sciences. 2026;16(11):5471. doi:10.3390/app16115471
BibTeX
@article{Tashenova2026,
author = {Tashenova, Zhuldyz and Aitmagambetuly, Aisultan and Bayegizova, Aigulim and Santeyeva, Saya and Abdugulova, Zhanat and Amanzholova, Shirin and Kerim, Akerke},
title = {SentinelCMS: Proactive Vulnerability Detection in CMS Plugins Using Static Taint Analysis and Bidirectional LSTM},
journal = {Applied Sciences},
year = {2026},
volume = {16},
number = {11},
pages = {5471},
doi = {10.3390/app16115471}
}
EUVIMED