purview-ai-hub

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

• Third-party content exposure detected (medium risk: 0.65). Third-party AI discovery and prompt evidence are ingested from runtime browser/endpoint DLP telemetry (Edge for Business + Defender for Cloud Apps / endpoint DLP agent), which can include free-form user-entered prompt text authored by outsiders (the end users) and/or third-party AI content, and that readable text is then used in the agent’s LLM context for analysis.