Skip to content

Security: KernFerm/MediaHarbor

docs/SECURITY.md

MediaHarbor Security

This document gives a short overview of MediaHarbor's current security posture and how to report a security issue.

Supported Version

  • 0.1.0

Security Scope

MediaHarbor is intended for lawful personal media downloads from public content only.

The project does not include:

  • DRM bypassing
  • paywall bypassing
  • private account access
  • login-only scraping
  • hidden telemetry or analytics

Core Protections

  • renderer nodeIntegration is disabled
  • contextIsolation and sandbox are enabled
  • external links are validated before opening
  • public URLs are validated before use
  • tunnel/backend URLs now include DNS-backed blocking for private, loopback, and link-local targets
  • filenames are sanitized before files are written
  • output paths are checked to stay inside the selected folder
  • raw user input is not passed directly into shell commands
  • sensitive settings are encrypted locally with a per-install secret
  • older local settings can be migrated forward to the stronger encryption model
  • download history stays on the local device

Reporting Security Issues

Please report security issues privately and avoid posting exploit details in public right away.

Include:

  • a short description
  • affected feature or file
  • reproduction steps
  • expected behavior
  • actual behavior

Notes

  • unsigned Windows installers may still trigger SmartScreen warnings
  • bundled tools like yt-dlp and ffmpeg should be kept up to date
  • 0.1.0 includes current bundled media tools and the current URL-validation and local-encryption hardening

There aren't any published security advisories