AMWD'22-阿里云安全WEBSHELL文本检测数据集

恶意程序/恶意代码攻防是云安全-主机安全领域中的一个关键领域，也是安全行业多年一直在攻关与竞争的必争之地。利用恶意程序/恶意代码，攻击者在Web服务器上执行系统命令、窃取数据、植入病毒、勒索核心数据、SEO挂马等恶意操作，危害极大。<br />
近几年随着攻防对抗不断升级，防御的挑战越来越大，在对抗的过程中，逐步发展出了静态检测引擎+AI检测引擎+动态沙箱执行检测引擎等多种综合手段，有效地提高了攻击者绕过的门槛和成本，缓解了恶意程序/恶意代码攻击问题。
<br />
本赛题数据就是在这样的背景下产生的，数据集来源于“AMWD 2022：阿里云安全WEBSHELL文本检测”挑战赛的训练集： https://tianchi.aliyun.com/competition/entrance/532035/introduction

Offense and defense against malware/malicious code constitutes a critical domain within cloud security and host security, and it has long been a fiercely contested strategic battleground that the cybersecurity industry has devoted years of efforts to research, address and compete for. By leveraging malware/malicious code, attackers can conduct various malicious operations on web servers, including executing system commands, stealing sensitive data, implanting viruses, ransoming core data, and planting malicious scripts for SEO tampering, inflicting extremely severe harm.

In recent years, as the offense-defense confrontation has continued to escalate, the challenges of defense have grown increasingly formidable. Amidst such confrontations, comprehensive defense measures integrating static detection engines, AI-powered detection engines, and dynamic sandbox execution detection engines have gradually been developed, effectively raising the threshold and cost for attackers to evade defenses and mitigating the threats posed by malware/malicious code attacks.

The dataset for this competition is developed against this backdrop, originating from the training set of the "AMWD 2022: Alibaba Cloud Security WEBSHELL Text Detection" challenge: https://tianchi.aliyun.com/competition/entrance/532035/introduction

搜集汇总

背景与挑战

背景概述

该数据集是阿里云安全在2022年发布的Webshell文本检测数据集，专门用于云安全领域的恶意程序检测。它包含2万多个样本，覆盖PHP和JSP两种语言，基于AST抽象语法树指令序列，旨在帮助研究人员构建二分类模型以识别Webshell文件，并支持静态检测与AI引擎的攻防对抗研究。