ai-visibility-audit

SUSPICIOUS. The stated purpose is coherent for an SEO/AI-visibility audit skill, and the visible skill text does not request disproportionate credentials or route data to unusual third-party endpoints. The main risks are transitive trust in delegated skills, missing command details in `references/checks.md`, and indirect prompt-injection exposure from analyzing arbitrary websites. No evidence here supports confirmed malware, but the incomplete execution context keeps risk above benign.