microsoft
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides functionality to read external web content from microsoft.com, which is an inherent surface for indirect prompt injection where instructions on a webpage could attempt to influence agent behavior. 1. Ingestion points: microsoft_com__read_page tool in SKILL.md. 2. Boundary markers: Absent from the instructions. 3. Capability inventory: The agent receives text and links which it may process or act upon. 4. Sanitization: The skill relies on the vendor's server-side proxy and the agent's internal safety filters.
- [EXTERNAL_DOWNLOADS]: The skill connects to the vendor's infrastructure for MCP communication and tool execution. Evidence: References to https://unbrowse.ai/mcp/microsoft.com in SKILL.md and server.json.
Audit Metadata