metatron-pentest-assistant

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The skill workflow involves scanning targets and searching external CVE/exploit databases via DuckDuckGo, where outsider-authored content (such as web pages or scan outputs from target systems) is fetched and passed to the LLM.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs cloning an arbitrary GitHub repository during setup, which introduces an unverifiable third-party dependency from an unofficial source.