azure-advisor
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest external documentation from learn.microsoft.com. While this creates an ingestion surface, the source is an official vendor domain.
- Ingestion points: Documentation is retrieved from learn.microsoft.com via mcp_microsoftdocs:microsoft_docs_fetch or fetch_webpage (SKILL.md).
- Boundary markers: The instructions do not specify explicit delimiters to isolate external content from the agent's internal instructions.
- Capability inventory: The skill's operations are restricted to network fetching and information retrieval; it does not utilize file writing, code execution, or system modification tools.
- Sanitization: No specific sanitization or validation steps are defined for the fetched Markdown content.
- [EXTERNAL_DOWNLOADS]: The skill fetches content from learn.microsoft.com and references the official MicrosoftDocs GitHub repository for tool documentation. These external resources are hosted on trusted vendor infrastructure.
Audit Metadata