azure-stream-analytics
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEOBFUSCATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [OBFUSCATION]: The skill contains a URL where the organization name is partially hidden using percent-encoding.
- Evidence:
https://github.com/Microsoft%44%6F%63%73/agent-skillsinSKILL.md. - Decoded:
https://github.com/MicrosoftDocs/agent-skills(targets a trusted vendor). - [INDIRECT_PROMPT_INJECTION]: The skill uses tools to fetch content from external documentation sites, which constitutes an attack surface.
- Attack Surface: Ingestion of untrusted data from the web via
mcp_microsoftdocs:microsoft_docs_fetchandfetch_webpagetools. - Potential Risk: Malicious instructions embedded in documentation could potentially influence agent behavior.
- [EXTERNAL_DOWNLOADS]: The skill references a large number of documentation resources from a well-known service.
- Source:
learn.microsoft.com(Microsoft). - Context: Used to provide technical guidance on Azure Stream Analytics.
Audit Metadata