azure-ai-contentsafety-py

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md includes an "Image from URL" example (AnalyzeImageOptions with ImageData(blob_url="https://...")) and text analysis/blocklist examples that clearly accept and fetch external URLs or user-provided text for analysis, meaning untrusted third‑party content is ingested and can materially influence decisions like halting on blocklist hits.