The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill includes extensive documentation for Azure CLI (az) commands to manage application registrations, service principals, and permissions. These are standard administrative tools, and the skill provides appropriate warnings regarding destructive operations like credential resets.
[EXTERNAL_DOWNLOADS] (SAFE): The skill references official Microsoft identity and security libraries (e.g., azure-identity, msal) from trusted package registries (PyPI, npm, NuGet). It also includes a Bicep template that pulls a Microsoft Graph extension from the trusted Microsoft Container Registry (mcr.microsoft.com).
[CREDENTIALS_UNSAFE] (SAFE): No real credentials or secrets are hardcoded. The skill uses clearly labeled placeholders (e.g., YOUR_APPLICATION_ID, your-secret-value-SAVE-THIS) and provides strong recommendations for using Azure Key Vault and Managed Identities in production environments.
[PROMPT_INJECTION] (SAFE): The content is strictly technical and instructional. It contains no attempts to override agent behavior, bypass safety filters, or extract system prompts.
[DATA_EXFILTRATION] (SAFE): All network-related references and OAuth endpoints point to official Microsoft domains (e.g., login.microsoftonline.com, graph.microsoft.com). No suspicious data exfiltration patterns or unknown external URLs were detected.

entra-app-registration