azure-architecture-autopilot

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

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly performs runtime web_fetch of MS Docs (e.g., to verify model availability via https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models) and requires those fetched pages to decide available models, API versions, SKUs and to generate Bicep code, so external content is fetched at runtime and directly controls the agent's prompts/choices and generated code.