Skip to content

Security: raphaelsalaja/pond

Security

SECURITY.md

Security Policy

Supported versions

Version Supported
0.0.x Yes

Reporting a vulnerability

If you discover a security issue in Pond, report it responsibly:

  1. Do not open a public GitHub issue for undisclosed vulnerabilities.
  2. Email or DM the maintainer via GitHub: @raphaelsalaja.
  3. Include steps to reproduce, impact, and affected version if known.

We aim to acknowledge reports within a few days and will coordinate disclosure once a fix is available.

Security model (summary)

Browser extension

The extension talks to Pond through Chrome native messaging (so.pond.host). The browser launches a small host binary that Pond registers; only this extension ID is allowed. On macOS, Pond writes a host manifest into each supported browser's NativeMessagingHosts folder. On Windows, Pond writes the manifests under its app data folder and registers them in per-user HKEY_CURRENT_USER registry keys, so no administrator rights are needed. The host reaches Pond over a Unix socket on macOS and a per-user named pipe on Windows. Saves and session cookies stay on your computer.

A loopback HTTP API on 127.0.0.1:41610 still exists for debugging. It requires the ingest token.

IPC

Renderer ↔ main IPC handlers validate the sending frame. Untrusted web content cannot invoke privileged channels.

Library data

Pond stores Library files, thumbnails, backups, and the search index unencrypted on your computer. Logs, preferences, and files you export are also plaintext. Keep full-disk encryption (FileVault on macOS, BitLocker on Windows) and standard account security on to protect them at rest.

Best practices for users

  • Rotate the ingest token if you suspect it was exposed (menu bar → Rotate Ingest Token).
  • Pond for Windows is not code-signed unless a release says otherwise. Download the installer only from official GitHub Releases and expect a SmartScreen prompt.
  • Only load the Pond extension from official GitHub Releases or a build you compiled yourself.

There aren't any published security advisories