azure-communication-callautomation-java

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFE
Full Analysis
  • [Authentication Practices]: The skill correctly demonstrates the use of DefaultAzureCredentialBuilder, which is the recommended approach for managing cloud identities without exposing secrets in code. It also includes warnings against hardcoding connection strings in its acceptance criteria.
  • [Network and Callback Security]: The skill utilizes callback URLs and media source URLs. All examples use placeholder domains (e.g., your-app.com) or official Azure infrastructure (azure.com), representing standard integration patterns for communication services.
  • [Data Ingestion Surface]: The skill implements webhook handlers that process incoming call events and speech/DTMF recognition results.
  • Ingestion points: Webhook request bodies in the handleCallback and handleCallEvents methods (SKILL.md, references/examples.md).
  • Boundary markers: None explicitly shown for the raw string data before parsing with CallAutomationEventParser.
  • Capability inventory: The skill uses capabilities like making calls (createCall), playing audio (play), and downloading recordings (downloadTo).
  • Sanitization: The skill uses the official CallAutomationEventParser to validate and convert raw JSON payloads into structured event objects, which is a standard safety measure for this service.
  • [Remote Package Management]: The skill references the official com.azure:azure-communication-callautomation package from the Microsoft Maven registry, which is an expected and verified dependency.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 03:16 PM