From Detection to Defense: A Comprehensive Benchmark and Study of Static Analysis Tools for PHP WebShell Detection
Scholarly Journal of Science and Technology Research and Development · 17.08.2026 · 5 (3) · 95-109
DOI: 10.46654/ac0w3647 ↗Kurz gefasst
Fachartikel in der Zeitschrift Scholarly Journal of Science and Technology Research and Development, online erschienen am 17.08.2026. Verfasst von Cyrus Ebere Orji. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
WebShells continue to be one of the most widespread and damaging threats to web infrastructure. PHP-based variations are responsible for the bulk of attacks due to the fact that PHP commands a market share of 73.6% within server-side programming languages. The absence of systematic benchmarking frameworks has made it difficult to conduct objective evaluations and meaningful comparisons of detection capabilities, despite the fact that various static analysis tools have been developed to identify malicious PHP scripts. Through the construction and evaluation of WebShellScanner, a unique entropy-aware detection framework, this academic paper presents a comprehensive benchmark research of static analysis tools for PHP WebShell detection. The goals of this study are to investigate the effectiveness of these tools. We measure the detection accuracy, false positive rates, and runtime performance of our tool by comparing it against a curated dataset consisting of 2,188 different WebShell samples and six PHP apps that are not malicious. The results of our experiments, which were carried out on a computer with an Intel Core i7 CPU, 12GB of RAM, WAMP Server 3.3.7 and PHP 8.3.4, show that WebShellScanner achieves a true positive rate of 97.2% while retaining a false positive rate of 4.1%, performing better than other open-source tools that are currently available. In this paper, we demonstrate that structural representations of code behaviour, when paired with severity-based categorization, significantly outperform standard signature-based techniques. In addition, we present an LLM-assisted regex generation strategy that is coupled with the VirusTotal API. This approach achieves a true positive rate of 92% while also retaining an incredibly low false positive rate of 0.49% at the same time. These findings not only establish a solid foundation for future research in automated WebShell identification and threat intelligence, but they also equip practitioners with insights that may be put into action directly.
Zitieren
APA
Orji, C. E. (2026). From Detection to Defense: A Comprehensive Benchmark and Study of Static Analysis Tools for PHP WebShell Detection. Scholarly Journal of Science and Technology Research and Development, 5(3), 95-109. https://doi.org/10.46654/ac0w3647
Vancouver
Orji CE. From Detection to Defense: A Comprehensive Benchmark and Study of Static Analysis Tools for PHP WebShell Detection. Scholarly Journal of Science and Technology Research and Development. 2026;5(3):95-109. doi:10.46654/ac0w3647
BibTeX
@article{Orji2026,
author = {Orji, Cyrus Ebere},
title = {From Detection to Defense: A Comprehensive Benchmark and Study of Static Analysis Tools for PHP WebShell Detection},
journal = {Scholarly Journal of Science and Technology Research and Development},
year = {2026},
volume = {5},
number = {3},
pages = {95--109},
doi = {10.46654/ac0w3647}
}
EUVIMED