Skip to content

Security: whosramoss/audiora

Security

SECURITY.md

Security Policy

Supported versions

Version Supported
1.x ✅

Reporting a vulnerability

If you discover a security issue, please do not open a public GitHub issue with exploit details.

Instead, use one of the following:

  1. GitHub Security Advisories (preferred) — private report on the repository.
  2. Contact the maintainer via the email listed on the npm package page if you cannot use GitHub.

Include:

  • A description of the issue and impact
  • Steps to reproduce
  • Affected versions
  • Any suggested fix (optional)

We aim to acknowledge reports within 7 days and will coordinate disclosure and a fix before publishing details when appropriate.

Scope

This library runs in the browser. It loads images, samples pixels via canvas, and generates audio with the Web Audio API (live and offline). Reports related to:

  • Cross-origin image handling / canvas tainting when string URLs are loaded
  • Unexpected network or credentialed requests triggered by library image loading
  • Denial of service from pathological images or parameter values that exhaust memory/CPU in library code paths (processing, scheduling, offline render)
  • Unsafe handling of consumer-supplied PresetData / render options that leads to script execution or unexpected privilege (when attributable to the library)

are in scope when they stem from library behavior, not from application misuse (e.g. embedding untrusted remote image URLs without CORS awareness, or granting the page unintended permissions).

Out of scope:

  • Browser autoplay / AudioContext resume policies
  • Vulnerabilities in the demo site hosting (www/), third-party CDNs, or consumer apps that embed audiora incorrectly
  • Hearing-safety / loudness of synthesized audio (callers control volume and should clamp for end users)
  • Issues that only affect Node or non-browser runtimes (the library targets browsers)

There aren't any published security advisories