dangling-markup-injection

Fail

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: CRITICALDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [DATA_EXFILTRATION]: The skill provides detailed instructions and ready-to-use payloads for exfiltrating sensitive information (CSRF tokens, API keys, session IDs) from a target web page. It explains how to use unclosed HTML tags to capture page content and send it to an external server controlled by an attacker.
  • [EXTERNAL_DOWNLOADS]: The documentation contains multiple references to "https://attacker.com/collect". This domain was flagged by automated security scanners as potentially malicious/phishing, though it is often used as a generic placeholder in security research examples.
Recommendations
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 9, 2026, 11:59 AM