The Agent Skills Directory

[SAFE]: The skill documentation follows security best practices for Azure CLI usage. It explicitly warns against insecure practices such as passing passwords as command-line arguments and committing service principal credentials.
[SAFE]: Includes a detailed 'Agent Safety Model' that classifies operations by risk (READ, WRITE, DESTRUCTIVE, EXPENSIVE, SECURITY, FORBIDDEN) and defines appropriate confirmation gates for the agent.
[SAFE]: Recommends secure authentication methods such as Managed Identities, Service Principals with certificates, and OIDC federated tokens for CI/CD environments.
[SAFE]: Demonstrates infrastructure-as-code deployment using Bicep and ARM templates from well-known services like GitHub, which are considered safe sources.
[SAFE]: Does not contain any code obfuscation, malicious persistence mechanisms, or unauthorized data exfiltration patterns.

azcli