fsi-ib-pitch-deck

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

• Third-party content exposure detected (high risk: 0.70). The workflow’s “Phase 1: Data Extraction” explicitly allows using “web sources” and “external documents” (public web content fetched at runtime / downloaded files from outside the workspace), which would be ingested as readable text into the agent’s LLM context during extraction/mapping.