Skip to content

Security: TimLai666/usegolib

Security

docs/security.md

Security Policy And Boundaries

This document describes what usegolib verifies, what it does not, and the expected trust boundaries.

Threat Model (In Scope)

  • Accidental corruption or tampering of built artifacts (shared libraries) on disk.
  • Unsafe archive extraction when bootstrapping build tools.

Threat Model (Out Of Scope)

  • Malicious Go modules: building and loading arbitrary native code is inherently unsafe. If you do not trust the source, do not build/load it.
  • Sandbox escape: artifacts run as native code inside the Python process.

Current Hardening

  • Artifact shared library integrity is verified against manifest.json SHA256 before loading.
  • Zig bootstrap downloads are restricted to https://ziglang.org/... and SHA256-verified using Zig's published metadata.
  • Zig archive extraction rejects absolute paths and path traversal entries.

Operator Guidance

  • Prefer shipping wheels containing prebuilt artifacts for end-users.
  • Pin module versions for production (avoid @latest).
  • Treat artifact roots as an integrity-sensitive directory (no untrusted writes).

There aren’t any published security advisories