research-studio

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md "Retrieval workflow" / "Default workflow" and related reference files like references/query-planning.md and references/query-examples.md) explicitly directs the agent to run initial scans and targeted searches of public sources (including platform examples such as 小红书 and public product/market commentary), synthesize and cite that third‑party content, and base conclusions and recommendations on it—therefore it clearly ingests untrusted public/user-generated content that can influence decisions.