Skip to content

Security: adamriofc/NeuronixOS

Security

SECURITY.md

Security Policy

Supported Versions

NEURONIX OS follows the active upstream NixOS channel lifecycle. Security updates and CVE mitigations are delivered through upstream nixpkgs and distribution modules.

Version Supported Maintenance Phase
1.0.x (Current) Yes Active Security, Testing, & Subsystem Hardening
< 1.0.0 No End of Life (Superseded)

Reporting a Vulnerability

The NEURONIX security team takes vulnerabilities seriously. If you discover a security flaw or vulnerability within NEURONIX OS (including installer generator scripts, system modules, or CLI utilities), please report it responsibly.

Disclosure Guidelines

  • Do not open a public issue. Please report security issues privately via GitHub Security Advisories at https://github.com/adamriofc/NeuronixOS/security/advisories/new.
  • Alternatively, report directly by email to: adamriofc@protonmail.com with the subject tag [SECURITY: NEURONIX].
  • Please provide:
    • Description of the vulnerability and its potential impact.
    • Minimal reproducible steps or proof-of-concept (PoC).
    • Target component (e.g., installer engine, PAM/PolKit configuration, memory shield, micro-VM isolation).
  • We will acknowledge receipt within 48 hours and coordinate a patch and CVE disclosure timeline.

Security Architecture & Threat Model

NEURONIX implements defensive engineering across its operational layers:

1. Immutable Store Protection

The entire package hierarchy (/nix/store) is mounted read-only at the kernel level. Package paths are derived from cryptographic SHA-256 hashes of their dependency closures. The immutable, hash-addressed store architecture prevents ordinary in-place mutation or tampering of existing store paths.

2. Privilege Separation & User Namespaces

  • Administrative actions require explicit authentication via PolKit or wheel-restricted sudo.
  • Non-root CLI operations (neuronix run, neuronix dev) execute inside unprivileged Linux user namespaces without requiring raw root access.

3. Isolated Sandbox Execution

The neuronix sandbox simulation framework (alias: neuronix try) runs untrusted or candidate system configurations inside an ephemeral QEMU micro-VM in RAM (/dev/shm). Host storage access is strictly restricted to read-only 9P filesystem mounts of /nix/store.

4. Processor & Kernel Hardening

  • Automatic CPU microcode updates are enabled by default for Intel and AMD platforms to mitigate transient execution flaws (Spectre, Meltdown, Zenbleed).
  • Secure Boot integration is supported via signed shim bootloaders utilizing lanzaboote.

There aren't any published security advisories