azure-keyvault-keys-ts

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFE
Full Analysis
  • Secure Authentication Patterns: The skill correctly implements DefaultAzureCredential for authentication, which is a best practice that automatically selects the most secure credential type based on the environment (e.g., Managed Identity in production, local CLI credentials in development).
  • Use of Official SDKs: The implementation relies exclusively on official Microsoft packages (@azure/keyvault-keys, @azure/keyvault-secrets, and @azure/identity), ensuring reliability and adherence to expected behavior.
  • Data Handling and Environment Management: Configuration is handled through environment variables (AZURE_KEYVAULT_NAME), avoiding hardcoded URLs or sensitive metadata within the skill files.
  • Indirect Prompt Injection Surface: Like any skill that retrieves data from external sources (in this case, Azure Key Vault secrets and key properties), there is a potential surface for indirect prompt injection if the retrieved content contains instructions. However, the skill treats this data as information for the user rather than instructions for the agent.
  • Ingestion points: Data is ingested through getSecret, getKey, and listing operations in SKILL.md, references/keys.md, and references/secrets.md.
  • Boundary markers: While explicit boundary markers are not shown in the snippets, the usage context is typically retrieving data for application configuration or cryptographic operations.
  • Capability inventory: The skill is capable of reading, writing, and deleting keys and secrets, as well as performing cryptographic operations (sign, encrypt).
  • Sanitization: The skill follows standard SDK practices, though users should ensure that data retrieved from secrets is handled appropriately by the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 07:40 AM