eu-ai-act-roles-oliver-schmidt-prietz

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

• Third-party content exposure detected (medium risk: 0.30). The skill ingests outsider-authored free text from the user at runtime in Phase 1 (the “Batch 1 — Open-ended question” asking the user to “answer in your own words”), and then internally extracts structured fields from that response without filtering for specific items.