Skip to content

Security: Vivswan/cloud-speech

SECURITY.md

Security policy

Supported versions

Only the latest release is supported.

Reporting a vulnerability

Do not open a public issue for security problems.

Report vulnerabilities privately via GitHub Security Advisories ("Report a vulnerability"). A useful report includes:

  • what an attacker can do (impact), and where trust is broken,
  • reproduction steps or a proof of concept,
  • the affected version or commit.

Expect an acknowledgement within a few days, and a fix in the next release once the report is confirmed. Please allow reasonable time for that fix before any public disclosure.

Never include real credentials in a report; redact everything that looks like a key.

Scope notes for researchers

One build is published to all three Chrome Web Store listing IDs (the unified Cloud Speech listing plus the two original fork listings) and to addons.mozilla.org, all at the same version; a report against any listing applies to all of them.

  • The extension stores user-provided API credentials (AWS, Azure, Google, OpenAI) in chrome.storage: sync by default, local when the user turns the sync toggle off. Anything that exfiltrates, logs, or leaks these credentials is in scope and high severity.
  • Selected page text is sent only to the TTS provider the user configured, directly from the browser, with no intermediary servers or analytics. Any destination for that text other than the four providers' official endpoints is a bug.
  • The content script runs on all pages (<all_urls>) to show error toasts; selected text is read on demand via scripting.executeScript. Injection or privilege-escalation findings in either path are in scope.

There aren't any published security advisories