xslt-injection
Fail
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPERSISTENCEOBFUSCATIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides numerous templates for remote code execution targeting various XSLT processors, including PHP's registerPHPFunctions, Java's Xalan/Saxon extensions (Runtime.exec), and .NET's msxsl:script block.
- [DATA_EXFILTRATION]: Instructions are provided for exfiltrating sensitive system files like /etc/passwd, /etc/shadow, and win.ini via document() and unparsed-text() functions, as well as Out-of-Band (OOB) data transfer to attacker-controlled domains.
- [COMMAND_EXECUTION]: Exploitation payloads explicitly use shell commands such as /bin/sh, cmd.exe, whoami, and id to demonstrate and achieve code execution.
- [EXTERNAL_DOWNLOADS]: The skill documents how to perform remote XSLT inclusion and SSRF via xsl:include and document() using URLs pointing to external attacker servers.
- [PERSISTENCE]: Includes instructions for establishing persistence by writing malicious payloads directly to system cron directories (/var/spool/cron/crontabs/www-data) using XSLT file-writing capabilities.
- [OBFUSCATION]: Provides specific techniques for WAF evasion, including the use of character entity encoding (e.g., a), CDATA wrappers, and dynamic string translation/concatenation to hide malicious keywords and paths.
- [DYNAMIC_EXECUTION]: Leverages advanced XSLT features like xsl:evaluate for dynamic XPath construction and execution, and C# scripting within .NET XSLT blocks.
Recommendations
- AI detected serious security threats
Audit Metadata