app-platform-sandbox

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides the do-app-sandbox SDK for executing arbitrary commands and scripts within isolated containers. This is the primary intended feature of the skill, specifically for supporting AI code interpreter workflows where isolation is required.
  • [EXTERNAL_DOWNLOADS]: Pulls pre-configured container images (sandbox-python and sandbox-node) from the GitHub Container Registry (ghcr.io/bikramkgupta) to provision the execution environments.
  • [EXTERNAL_DOWNLOADS]: Instructions include installing the do-app-sandbox package from public Python package registries to manage the sandbox lifecycle and file operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 01:57 PM
Security Audit — agent-trust-hub — app-platform-sandbox