agent-framework-azure-ai-py

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

• Third-party content exposure detected (low risk: 0.10). SKILL.md describes agent runtime as taking a user query string passed directly to agent.run(...)/agent.run_stream(...) and configuring tools (e.g., HostedWebSearchTool), but it does not show any required workflow that ingests outsider-authored free text from an external feed/queue without the agent first selecting a specific item.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The example creates a MCPStreamableHTTPTool with url="https://learn.microsoft.com/api/mcp" which is used at runtime to stream client-managed MCP content into the agent and therefore can directly control prompts/instructions.