mixed-content-scan

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

• Third-party content exposure detected (high risk: 0.85). The skill navigates to a user-supplied HTTPS URL and then ingests the page’s DOM-derived strings (e.g., document.scripts, link[href], document.images, iframe[src], etc.) via Playwright into the LLM context, which can include outsider-authored free text/URLs from that fetched web page (public web content at runtime).