azure-resource-manager

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow explicitly fetches remote Microsoft Learn documentation via mcp_microsoftdocs:microsoft_docs_fetch (or WebFetch fallback), which ingests outsider-authored free-form article text into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires network access and instructs the agent to fetch Microsoft Learn articles at runtime (via mcp_microsoftdocs:microsoft_docs_fetch or WebFetch), for example https://learn.microsoft.com/en-us/azure/azure-resource-manager/bicep/bicep-core-diagnostics, and those fetched documents are injected into the agent context to drive its responses and thus directly control prompts.