entra-app-registration

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill guides the user to install official Microsoft packages such as 'msal', '@azure/identity', and 'azure-identity' from public registries. These are trusted, well-known dependencies for identity management.
  • [COMMAND_EXECUTION]: Multiple Azure CLI commands ('az ad') are provided for administrative tasks like creating app registrations and resetting credentials. These commands are appropriate for the skill's stated purpose.
  • [DATA_EXFILTRATION]: The provided code examples include network requests to official Microsoft Graph endpoints ('graph.microsoft.com'). These operations are necessary for the skill's functionality and target a trusted domain.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface where user-provided data is intended to be used in shell commands and code templates.
  • Ingestion points: Placeholders for application names, client IDs, and tenant IDs in files such as 'references/cli-commands.md' and 'references/first-app-registration.md'.
  • Boundary markers: There are no explicit delimiters or specific instructions for the agent to ignore potentially malicious embedded content within user-provided strings.
  • Capability inventory: The skill enables the execution of infrastructure-modifying CLI commands and the implementation of authentication flows using the provided examples.
  • Sanitization: No input validation or escaping mechanisms are described to protect against command injection via the user-supplied variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 12:34 AM