azsdk-common-apiview-feedback-resolution

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

• Third-party content exposure detected (high risk: 0.85). The required workflow calls azure-sdk-mcp:azsdk_apiview_get_comments, which retrieves APIView reviewer comments from a user-supplied APIView URL (public web content / outsider-authored text) and feeds that free-form comment text into the agent for categorization and resolution.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes the runtime tool azsdk_apiview_get_comments to fetch APIView pages (e.g., https://spa.apiview.dev/... and https://apiviewstagingtest.com/...) and injects the retrieved reviewer comments into the agent's context to drive prompt-based fixes, so the external URLs directly control agent instructions at runtime.