portal-backend-app

Pass

Audited by Gen Agent Trust Hub on Aug 23, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill references standard development and deployment commands (e.g., bun run dev:all, docker-compose) and configuration for the Caddy web server. These instructions are consistent with the monorepo's operational requirements and are for documentation purposes.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: Contains references to internal service endpoints and development-specific user accounts (e.g., rockie, liming, chenjing). These are identified as part of the vendor's (xgent-ai) testing and development infrastructure and do not include sensitive credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a framework for handling external authentication tokens (TDT). It explicitly prescribes security best practices to mitigate injection risks, including mandatory server-side token introspection and the enforcement of tenant isolation by relying on verified claims rather than request body data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 23, 2026, 08:41 AM
Security Audit — agent-trust-hub — portal-backend-app