microsoft-foundry
Pass
Audited by Gen Agent Trust Hub on May 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's instructions and scripts align with documented Azure administrative workflows. No malicious patterns such as prompt injection, credential theft, or unauthorized data exfiltration were found.
- [EXTERNAL_DOWNLOADS]: The skill downloads configuration templates and agent samples from trusted Microsoft-owned GitHub repositories (e.g., microsoft-foundry/foundry-samples, Azure-Samples/azd-ai-starter-basic). These resources are used to bootstrap agent projects and do not present a security risk.
- [COMMAND_EXECUTION]: Extensive use of the Azure CLI (az) and PowerShell/Bash scripts to manage cloud infrastructure, deployments, and telemetry. These operations are required for the skill's primary functions and are performed using standard, verifiable commands.
- [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes untrusted data such as agent responses and production traces (ingestion points in analyze-results.md and trace-to-dataset.md). However, it mitigates risk by recommending Managed Identity and providing structured analysis workflows. Capability inventory includes powerful tools like agent_update and monitor_resource_log_query. Sanitization of ingested trace data is not explicitly implemented in the provided scripts.
Audit Metadata