Skip to content

Security: leduckhc/makit

Security

SECURITY.md

Security Policy

Makit lets you drive coding agents (pi / codex / claude-code) on your desktop from a paired phone or other device. A compromise could give an attacker the ability to run those agents — and therefore code — on your machine, plus read access to whatever those agents can see. We take security seriously.

Supported versions

Makit is pre-1.0 and moves fast. Security fixes land on main and the latest release only.

Version Supported
latest main / release
older releases

Reporting a vulnerability

Please do not open a public GitHub issue for security vulnerabilities.

Report privately using either:

Please include:

  • A description of the issue and its impact.
  • Steps to reproduce (proof-of-concept if possible).
  • Affected component (app/, server/, pairing/transport, an adapter, …) and version / commit.

What to expect

  • Acknowledgement: within 3 business days.
  • Assessment & triage: we'll confirm the issue and share a rough timeline.
  • Fix & disclosure: we aim to ship a fix and publish an advisory promptly, crediting you unless you prefer to remain anonymous. Please give us a reasonable window to fix before any public disclosure (coordinated disclosure).

Scope

The security model in short (see app/SECURITY.md and server/SECURITY.md for details):

  • Private by default: the server binds to a Tailscale tailnet, not open Wi-Fi. Plain-LAN is opt-in (--lan).
  • Cert-pinned transport: the app pins the server's self-signed cert fingerprint at the WebSocket layer — no OS trust store involved.
  • Pairing: one-time QR pair token → long-lived bearer stored in the OS secure enclave (Keychain / Android Keystore).

In-scope reports include: authentication/pairing bypass, transport downgrade or cert-pinning bypass, bearer-token leakage, remote code execution via the agent adapters, and supply-chain issues in our dependencies.

Out of scope: vulnerabilities in the underlying agents (pi, codex, claude-code) themselves, in Tailscale, or in your own network configuration.

There aren't any published security advisories