topic-research

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

• Third-party content exposure detected (high risk: 1.00). The skill's Step 2: Research explicitly instructs the agent to perform thorough web research and "collect source URLs" from open/public websites (see the "Research Sources" section and example URLs like earthweb.com, mdpi.com, pmc.ncbi.nlm.nih.gov), so the agent will fetch and read untrusted third‑party web content as part of its workflow.