API Catalog

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No malicious instructions or bypass attempts were found. The skill includes a standard activation instruction directing the agent to use the provided references when users mention specific services.
  • [DATA_EXFILTRATION]: The skill emphasizes the 'Iron Law' of not hardcoding credentials, directing users to use Home Assistant's !secret tag or Node-RED environment variables. All network communication templates target legitimate, well-known API endpoints for functional purposes (e.g., fetching weather or transport data).
  • [COMMAND_EXECUTION]: No dangerous shell commands or privilege escalation attempts were detected. The skill focuses on standard HTTP/REST/GraphQL interactions within the Home Assistant and Node-RED ecosystems.
  • [EXTERNAL_DOWNLOADS]: The skill mentions and provides links to official developer documentation and well-known, trusted GitHub repositories for community integrations. No suspicious external scripts or binaries are downloaded.
  • [SAFE]: The code snippets provided are standard templates for home automation tasks. The skill's behavior aligns with its stated purpose of being an API reference guide.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 12:55 AM
Security Audit — agent-trust-hub — API Catalog