Security

Reporting a vulnerability

Use GitHub's private vulnerability reporting rather than a public issue. If that is unavailable, open an issue saying only that you have a security report and asking for a contact route. Do not put the details in a public issue.

Expect an acknowledgement within a week.

What is in scope

This is a local command line tool. It has no server, no network listener, and no authentication. The realistic attack surface is small but not empty:

• Path handling. The tool takes file paths from a CSV and from command line flags, and it deletes recursively when rebuilding an export. --export-dir refuses to remove files it did not create unless --force is passed. A way to make it delete something outside the directory it was pointed at is a bug worth reporting.
• Subprocess invocation. Every ffmpeg, Demucs, and TTS call goes through waze_voice/media.py as an argument list, never a shell string. A path or a config value that escapes into shell interpretation would be a real finding.
• Config and inventory parsing. config/*.json and the source CSV are parsed with the standard library. Unknown keys are rejected rather than ignored. Filenames in the inventory are required to be bare names, not paths, so a traversal through filename or waze_filename would be a bug.
• Optional dependencies. Demucs and the TTS backends download model weights from third-party hosts on first use. Those downloads are not verified by this project beyond whatever the libraries do themselves. Treat model weights the way you would treat any other third-party binary.

What is out of scope

• The behaviour of Waze, its servers, or its share links. This project does not control any of that and reports about it belong to Waze.
• The community upload tooling at pipeeeeees/waze-voicepack-links. Report those to that project.
• Audio you supply. If you point the tool at a malicious media file, it is ffmpeg that parses it; report parser bugs upstream to FFmpeg.
• Rights and licensing questions. Those are real, but they are not security issues. See LEGAL.md.

A note on what this tool does

It clones voices and it publishes packs to a third-party service when you upload one. Neither is a vulnerability, but both are worth being deliberate about. The synthesis step asks you to acknowledge that you have the rights and consent for the voice you are cloning, and the export step tells you plainly that uploading publishes the result behind a shareable link.