building-inferencesh-apps

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions for installing the platform's command-line interface and other development dependencies using shell-piped download commands from official vendor and well-known service domains.
  • [EXTERNAL_DOWNLOADS]: Fetches installation scripts and configuration from the vendor's infrastructure and established developer service providers.
  • [COMMAND_EXECUTION]: The instructions involve executing various lifecycle commands through the infsh CLI to scaffold, test, and deploy applications.
  • [PROMPT_INJECTION]: The application templates define input schemas that process external data; while these are standard for application development, the examples do not explicitly demonstrate sanitization techniques for untrusted user content. • Ingestion points: AppInput (Python) and RunInput (Node.js) schemas. • Boundary markers: Absent in boilerplate templates. • Capability inventory: Apps can perform file writes, network operations, and execution via platform kernels. • Sanitization: No explicit sanitization or validation logic shown in templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 02:29 PM