survey

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 (Execution Policy: "Search broadly in English", "Keep claims source-backed. Include links") and the Platform Adapter / Lane A instructions explicitly require using web-capable agents and collecting "direct user voices from communities", so the agent will fetch and read untrusted public web/forum/social content as part of its research and use those findings to shape decisions.