azure-compliance
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the Azure environment (resource names, tags, and audit reports) and possesses tools capable of performing sensitive operations.\n
- Ingestion points: Processes Azure resource inventory, names, tags, and azqr scan results (Excel/JSON/Markdown) as described in
references/azure-quick-review.md.\n - Capability inventory: Accesses Key Vault secrets, keys, and certificates via
keyvault_*tools, and includes instructions for shell command execution using the Azure CLI fallback as documented inreferences/azure-keyvault-expiration-audit.md.\n - Boundary markers: No explicit instructions or delimiters are used to warn the agent to ignore instructions embedded in the resource data.\n
- Sanitization: No sanitization or validation of the ingested resource metadata is performed before processing.\n- [COMMAND_EXECUTION]: The skill contains multiple Azure CLI, Bicep, and PowerShell code snippets intended for manual or agent-assisted remediation of compliance findings. Evidence found in
references/azqr-remediation-patterns.mdshowing scripts for storage, Key Vault, and AKS remediation.
Audit Metadata