agent-framework-azure-ai-py

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md examples and tool list explicitly include a HostedWebSearchTool (Bing) and MCPStreamableHTTPTool for fetching web content (e.g., the "HostedWebSearchTool(name='Bing')" and the "Search for Python best practices and summarize" example), so the agent will ingest and act on untrusted public web pages that could contain instructions influencing its behavior.