microsandbox
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides tools for secure code isolation using hardware-level microVMs. It includes explicit security warnings to treat sandbox output as untrusted data rather than instructions, and to use least-privilege configurations for network and filesystem access.
- [EXTERNAL_DOWNLOADS]: The skill references installation commands for the microsandbox CLI and SDKs via standard registries (Homebrew, npm, uv, cargo, pip, go). These resources are official packages from the vendor 'superradcompany'.
- [CREDENTIALS_UNSAFE]: The skill includes robust guidance on secret management, advocating for the use of host environment variables and placeholder substitution to ensure real credentials never enter the untrusted sandbox environment, effectively mitigating credential exposure risks.
Audit Metadata