Skip to content

Security: 0xbrando/dictate

Security

SECURITY.md

Security Policy

Supported Versions

Version Supported
2.5.x
2.4.x
< 2.4

Architecture

Dictate is designed with security as a first principle:

  • 100% local processing — no audio or text leaves your machine
  • No network calls except to localhost LLM endpoints and optional update checks
  • Model repository allowlist — only mlx-community/ prefixed models accepted
  • No telemetry, no analytics, no tracking
  • Clipboard access is the only system integration (required for auto-paste)

Reporting a Vulnerability

If you discover a security vulnerability, please report it responsibly:

  1. Do NOT open a public issue
  2. Open a private security advisory on GitHub
  3. Include:
    • Description of the vulnerability
    • Steps to reproduce
    • Potential impact
    • Suggested fix (if any)

We will acknowledge receipt within 48 hours and aim to release a fix within 7 days for critical issues.

Past Security Fixes

  • v2.3.1 — Fixed 6 medium-severity issues: endpoint validation hardening, plist generation security, dependency pinning
  • v2.0.0 — Model repository allowlist, input sanitization

Scope

The following are in scope:

  • Code execution via crafted input
  • Model loading from unauthorized sources
  • Clipboard data leaks
  • Privilege escalation

The following are out of scope:

  • Physical access attacks
  • Accessibility permission abuse (macOS system-level)
  • Attacks requiring modifying the installed package

There aren’t any published security advisories