google-ads-landing

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

• Third-party content exposure detected (high risk: 1.00). The skill's required workflow explicitly WebFetches and parses arbitrary landing-page URLs (Phase 2 "WebFetch the landing page" and the rule "Never score a page without WebFetch'ing it") — user-supplied or ad-discovered public pages — and then uses that page content to compute message-match scores and recommend actions, so untrusted third-party content can materially influence the agent's decisions.