# Contributing

Thanks for helping maintain this list.

## Adding a Paper

Please add each paper to one primary category file under `categories/`.

Recommended format:

```markdown
## Attack

| Year | Venue | Paper | Tags |
| --- | --- | --- | --- |
| 2026 | Venue | Paper Title | attack, keyword |
```

Keep category tables sorted newest first. Use one broad section: `Attack`, `Defense`, `Benchmark & Measurement`, or `Other`.

Detection and attribution papers belong under `Defense`. Include only high-quality main-conference papers or clearly strong arXiv preprints.

Datasets, benchmarks, surveys, evaluations, analyses, and measurement papers belong under `Benchmark & Measurement`. Use `Other` only for valuable papers that do not fit the first three sections.

Do not include short papers, Findings papers, workshops, demos, tutorials, tooling-only papers, position papers, or weak arXiv preprints by default.

If a paper fits multiple areas, choose the most specific primary category and add secondary tags.

## Category Guide

- Audio Deepfake, Voice Cloning, and Spoofing
- Voice Privacy, Anonymization, and Speaker Protection
- Watermarking, Provenance, and Data Rights
- ASR and Speech Translation Security
- Audio-Language Model Safety
- Side Channels and Physical Eavesdropping
- Music and Singing Voice Security
- Voice Authentication and Biometrics

## Summary Style

- Keep summaries factual and short.
- Avoid overstating security guarantees.
- Mention the threat model when it is central to the paper.
- Prefer official paper, code, dataset, or project links.
- Do not include private notes or unreleased experiment details.