azure-enterprise-infra-planner
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFE
Full Analysis
- Structured Deployment Workflow: The skill implements a rigorous six-phase lifecycle (Research, Refine, Plan, Verify, Generate, Deploy) to ensure infrastructure is planned before deployment. It uses a JSON-based plan schema to maintain state and requires a status transition to 'approved' before execution, serving as a significant safety checkpoint.
- Security Best Practices Integration: The instructions consistently prioritize security hardening, such as using Managed Identities over connection strings, enabling Key Vault soft-delete/purge protection, and disabling local authentication on data services. It explicitly follows the Azure Well-Architected Framework (WAF) pillars.
- Authoritative Guidance Pattern: The SKILL.md file contains strong instructional language (e.g., 'AUTHORITATIVE GUIDANCE', 'MANDATORY COMPLIANCE') designed to ensure the model follows the provided technical constraints and local resource references rather than its training data. While similar to some prompt injection patterns, in this context it is used constructively to prevent architectural hallucinations and ensure compliance with specific Azure naming and pairing rules.
- Safe CLI and Tool Usage: The skill utilizes standard industry tools including the Azure CLI (
az) and Terraform. It incorporates syntax validation steps (e.g.,az bicep build,terraform validate) and 'what-if'/preview analysis before applying changes to live environments. - Trusted Resource References: All external URLs target official Microsoft documentation (
learn.microsoft.com) and well-known registries, which are considered safe sources. No unauthorized external network operations or data exfiltration patterns were observed.
Audit Metadata