project

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflows explicitly call for external web/public-data ingestion — e.g., evaluation-protocol.md cites "AI 기반 검증: 사실 확인 (웹 검색)" and init-protocol.md / Phase 6 lists public APIs (data.go.kr, KIPRIS, law.go.kr, fal.ai, ElevenLabs) to be used — meaning the agent will fetch and interpret untrusted public third‑party content that can influence verification, decisioning, and subsequent tool actions.