microsoft-foundry-classic
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides numerous links to technical documentation hosted on
learn.microsoft.com, a well-known and trusted service. It instructs the agent to fetch content from these URLs for user queries. - [EXTERNAL_DOWNLOADS]: The instructions suggest the installation of an MCP (Model Context Protocol) tool from a repository under the
MicrosoftDocsGitHub organization, which is a trusted organization. - [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting data from external web pages. However, the ingestion is restricted to official Microsoft documentation domains, and the skill does not exhibit exploitable capabilities that would elevate this risk above a safe level.
- [SAFE]: The conditional logic regarding the
generated_atmetadata field is a benign maintenance check used to ensure the agent provides the user with up-to-date documentation links.
Audit Metadata