didit

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md AML Screening endpoint includes an include_adverse_media option that causes the API to ingest public adverse media (open web/news/public sites) as part of screening, and those external media hits are used to influence screening outcomes/approve-decline decisions, so untrusted third‑party content can affect agent behavior.