skills/vercel/vercel-plugin/sandbox/Gen Agent Trust Hub

sandbox

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFE
Full Analysis
  • [Sandbox Environment Management]: The skill documents the use of the @vercel/sandbox SDK to create and manage isolated execution environments. This includes patterns for persistent storage, resource allocation, and region-specific deployment, all of which are standard features for providing secure, isolated compute.
  • [Network Policy Configuration]: Instructions are provided for defining granular network access controls, including domain whitelisting and credential brokering. These features allow for the restriction of outbound traffic from within the sandbox, which is a recommended security practice for running untrusted code.
  • [Command Execution in Isolation]: The skill describes how to execute shell commands and manage background processes within the MicroVM. While this involves sudo usage and command execution, these operations are confined to the isolated guest environment and do not affect the host system or the agent's primary operating context.
  • [Scoped User Access]: The documentation includes patterns for creating isolated Linux users and groups within the sandbox to enforce internal permission boundaries, which supports multi-tenant or multi-agent workflows within a single VM instance.
  • [Official Resource Integration]: The skill references official documentation, CLI tools, and GitHub repositories from the vendor's verified domains and organizations, ensuring the agent uses legitimate and up-to-date integration patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 02:44 AM