azure-lab-services

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires runtime fetching of external Markdown documentation (via the mcp_microsoftdocs:microsoft_docs_fetch tool or fetch_webpage) from learn.microsoft.com (e.g., https://learn.microsoft.com) and points to the mcp GitHub install guide (https://github.com/MicrosoftDocs/mcp/blob/main/README.md), and those fetched documents are injected into the agent context and can directly influence prompts/instructions.