add-ai-webapi

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path for Phase 5/7 ingests outsider-authored free text into the LLM context via the Power Pages generative-AI summarization endpoints (POST /_api/search/v1.0/summary and POST /_api/summarization/data/v1.0/...), where the model grounds on site knowledge/content and record fields that are not authored by the operating user (e.g., knowledge articles, case comments, Dataverse record text), and that content is sent to the Azure OpenAI layer as part of the summarization request/response flow.