marketing-seo

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

• Third-party content exposure detected (high risk: 1.00). The skill's Research Mode explicitly instructs the agent to use agent-browser to open and scrape live public sites (e.g., "https://www.google.com/search?q={target-keyword}", SimilarWeb, pagespeed.web.dev, validator.schema.org, answerthepublic.com") and to "agent-browser get text body" and extract SERP/metrics which the agent then uses to drive recommendations and next actions, so it clearly ingests untrusted third‑party user-generated/public web content that can influence behavior.