entra-app-registration
Warn
Audited by Gen Agent Trust Hub on May 3, 2026
Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: Deceptive metadata regarding the skill's origin. The 'author' field is set to 'Microsoft' in the frontmatter, which contradicts the identified author 'attentiondotnet'. This mismatch can lead to user confusion regarding the official nature of the instructions.
- [EXTERNAL_DOWNLOADS]: Fetches configuration and official SDKs from trusted registries. The skill downloads the Microsoft Graph Bicep extension from the official container registry (mcr.microsoft.com) and references standard libraries like 'msal' and 'azure-identity' from public registries.
- [PROMPT_INJECTION]: Surface for indirect prompt injection within documentation templates.
- Ingestion points: User-provided inputs such as 'Application name' and 'Redirect URIs' are interpolated into CLI and code templates in 'references/cli-commands.md' and 'references/console-app-example.md'.
- Boundary markers: None identified in the provided templates.
- Capability inventory: The skill facilitates the execution of Azure CLI commands and MSAL-based authentication logic.
- Sanitization: No validation or escaping is applied to user-supplied variables within the provided examples.
Audit Metadata