azure-compute

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [External Data Retrieval]: The skill uses web_fetch to retrieve live documentation from learn.microsoft.com and curl or PowerShell to query the Azure Retail Prices API at prices.azure.com. These operations target official, trusted Microsoft domains to ensure recommendations are based on current data.
  • [Command Execution]: The skill instructions include examples of command-line tools like curl and Invoke-RestMethod to interact with the Azure Retail Prices API. These commands are used for data ingestion from a trusted vendor service and do not involve executing external scripts or binaries.
  • [Indirect Prompt Injection Surface]: Information is ingested from external web pages and API responses. Ingestion points: learn.microsoft.com and prices.azure.com. Boundary markers: The skill instructs the agent to verify requirements and uses fetched data for recommendation context. Capability inventory: web_fetch, curl, and Invoke-RestMethod. Sanitization: Not explicitly defined in the instructions, but the risk is mitigated by the use of high-reputation vendor sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 09:37 PM