Description
### Summary
Versions of `i18nextify` prior to 4.0.8 substitute `{{key}}` interpolation tokens inside `src` and `href` attribute values with the raw string returned by `i18next.t()`. The substitution logic in `src/localize.js` (`replaceInside` handler around line 122) only guards against a duplicated `http://` origin prefix — it does not validate the URL scheme of the substituted value. A translated value such as `javascript:alert(1)` or `data:text/html,<script>...</script>` is applied unchanged to the live DOM attribute.
### Impact
When an attacker can influence the content of a translation file or the translation-backend response — compromised translation CDN, user-contributed locales, MITM on a plain-HTTP backend, write access to the translation JSON — they can:
- Set any `href` on an anchor to a `javascript:` URI, executing arbitrary JavaScript when the victim clicks the link.
- Set any `src` on `<iframe>`, `<object>`, or `<embed>` to a `data:text/html` URI containing a full script payload that runs in the page's origin.
- Use `vbscript:` on legacy IE installations or `file:` for local-resource navigation attacks.
This path is distinct from the general i18nextify design that intentionally renders HTML from translations — href/src schemes are narrow and attack-specific, and no legitimate translation needs `javascript:` or `data:`. The fix therefore blocks these schemes outright without changing other behaviour.
### Also fixed in 4.0.8
- **`debug` / `saveMissing` URL-parameter substring match.** The previous detection `window.location.search.indexOf('debug=true') > -1` matched the substring anywhere in the query string. A URL like `?nosaveMissing=true` silently enabled `saveMissing` mode, causing the victim's browser to POST every unknown translation key to the configured `addPath` — a form of CSRF-style abuse of missing-key reporting. `?track_debug=true` enabled verbose debug logging, leaking i18next internals to the console. Now uses `URLSearchParams` for exact parameter matching.
- **Optional `sanitize(html, ctx)` hook.** The library's core purpose is to render HTML from translations — a behaviour that is safe only when the translation source is fully trusted. Applications with partially-trusted sources (user-contributed locales, third-party CDN, MITM-exposed HTTP backend) can now wire a sanitizer (e.g. DOMPurify) via `i18next.options.sanitize`. Defaults to pass-through to preserve existing behaviour for the main use case.
### Affected versions
All versions of `i18nextify` prior to **4.0.8**.
### Patch
Fixed in **4.0.8**. The URL-scheme blocklist is `^\s*(javascript|data|vbscript|file)\s*:` (case-insensitive) applied to each translated value before it is joined back into the `href`/`src` attribute. Values matching the blocklist are replaced with an empty string so the attribute becomes harmless rather than leaving the attacker's URL in place.
### Workarounds
No workaround short of upgrading. If you cannot upgrade immediately, audit every translation file for `javascript:`, `data:`, `vbscript:`, and `file:` prefixes in any value that may reach an `href`/`src` position, and restrict translation-file write access to trusted operators. Serving translations over HTTPS and pinning the translation backend to an internal origin reduce the MITM surface.
### Credits
Discovered via an internal security audit of the i18next ecosystem.