The Agent Skills Directory

Instructional Alignment Directives: The skill uses definitive language (e.g., "AUTHORITATIVE GUIDANCE", "supersedes all other sources") to ensure the AI agent follows the prescribed architectural workflow accurately. While this technique is used to maintain consistency and compliance with Azure best practices, it aligns with patterns designed to prioritize specific instructions over general model behavior.
Indirect Prompt Injection Surface: As an infrastructure planner, the skill translates user-provided goals into technical designs and executable code. This introduces a surface where user inputs could influence the resulting architecture. The skill addresses this potential consideration by implementing a sequential workflow with research, planning, and verification phases that require user approval before proceeding.
Dynamic Infrastructure-as-Code Generation: The skill generates Bicep and Terraform files based on live architectural requirements. This dynamic generation and subsequent execution via CLI tools (like az and terraform) are central to its function. To ensure reliability, the skill includes mandatory validation checks and destructive action confirmations.
High-Privilege Capability Usage: The skill facilitates subscription-level and resource-group-level deployments. These high-privilege operations are appropriate for enterprise infrastructure planning and are managed through structured status gates (e.g., requiring a status of "approved") and explicit user consent before execution.

azure-enterprise-infra-planner