GHSA-Q89C-Q3H5-W34G

Vulnerability from github – Published: 2026-04-22 17:41 – Updated: 2026-05-11 13:29
VLAI
Summary
i18next-http-backend has Path Traversal & URL Injection via Unsanitised lng/ns
Details

Summary

Versions of i18next-http-backend prior to 3.0.5 interpolate the lng and ns values directly into the configured loadPath / addPath URL template without any encoding, validation, or path sanitisation. When an application exposes the language-code selection to user-controlled input (the default — i18next-browser-languagedetector reads ?lng= query params, cookies, localStorage, and request headers), an attacker can inject characters that change the structure of the outgoing request URL.

Affected call sites:

  • _readAnylib/index.js:64: interpolate(resolvedLoadPath, { lng: languages.join('+'), ns: namespaces.join('+') })
  • createlib/index.js:123 (pre-patch): interpolate(addPath, { lng, ns: namespace })

The helper interpolate (lib/utils.js) previously returned the raw value with no encoding. In contrast, addQueryString already correctly uses encodeURIComponent for each query-string param — only the URL-path substitution was unprotected.

Impact

An attacker who can influence the resolved lng or ns value can alter the URL in several ways:

  • Path traversallng = '../../config' turns /locales/{{lng}}/{{ns}}.json into /locales/../../config/translation.json. On a misconfigured web server, this can cause the request to target a different resource than intended; in SSR pipelines that use file:// or similar schemes for loadPath, it can read arbitrary files from the host filesystem.
  • Query-string injectionlng = 'en?admin=true' turns /locales/{{lng}}/{{ns}}.json into /locales/en?admin=true/translation.json. Some server frameworks parse the query portion with higher priority than the path and branch on attacker-controlled flags.
  • Fragment truncationlng = 'en#anything' silently discards the rest of the path in browser fetches (client cannot see the final URL).
  • URL-encoded bypasseslng = 'en%2F..', after server-side URL decoding, resolves to en/.. — the attacker bypasses the absence of a literal / in their input.

The practical worst case is SSRF when loadPath is an internal or file-scheme URL, and path-based authorisation bypass against servers that segment access by URL prefix.

Also fixed in 3.0.5

  • Per-instance omitFetchOptions. A module-level boolean in lib/request.js was flipped to true the first time any backend instance hit a "not implemented" fetch error. Once flipped, all subsequent requests from all backend instances in the same module silently stripped every user-configured fetch option — including security-relevant credentials, mode, and cache. One misbehaving instance (for example during SSR hydration or in React Native) permanently removed these protections process-wide. 3.0.5 scopes the flag to the backend's options object (options._omitFetchOptions) so one instance's fallback cannot pollute siblings.
  • Log forging via control characters in lng/ns. Error callbacks embedded the raw lng/ns/URL in the message string. Crafted CR/LF values could inject fake log lines into file-backed log aggregators (CWE-117). 3.0.5 strips C0/C1 control chars before concatenation.
  • Basic-auth credentials leaked into error callbacks. If loadPath contained a user:password@host authority, the full URL (including the credentials) ended up in the error message strings returned to the caller. 3.0.5 redacts user:password@ before logging.
  • Prototype-pollution amplification via for...in. addQueryString and the XHR customHeaders loop used for...in which walks the prototype chain. Polluted Object.prototype entries could leak into URL query parameters and request headers. 3.0.5 uses Object.keys and an explicit prototype-key guard.

Affected versions

All versions of i18next-http-backend prior to 3.0.5.

Patch

Fixed in 3.0.5. Summary of the hardening:

  1. New utils.interpolateUrl (used by _readAny and create) returns null if any substitution fails the URL-segment safety check (blocks .., /, \, ?, #, %, @, whitespace, control chars, prototype keys, and values > 128 chars). Multi-language joins (en+de) are validated per-segment. The call sites now refuse to issue a request when the check fails and call back with a clear error.
  2. omitFetchOptions is stored per-instance on options._omitFetchOptions.
  3. Error-callback messages sanitise strings and redact URL credentials.
  4. for...in over untrusted objects replaced with Object.keys + prototype-key guard.

Workarounds

No workaround short of upgrading. If you cannot upgrade immediately, sanitise lng / ns yourself before they reach i18next (strip .., /, \, ?, #, %, whitespace, and control characters; cap the length).

Credits

Discovered via an internal security audit of the i18next ecosystem.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "i18next-http-backend"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "3.0.5"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-41691"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22",
      "CWE-74"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-04-22T17:41:24Z",
    "nvd_published_at": "2026-05-07T21:16:29Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\n\nVersions of `i18next-http-backend` prior to 3.0.5 interpolate the `lng` and `ns` values directly into the configured `loadPath` / `addPath` URL template without any encoding, validation, or path sanitisation. When an application exposes the language-code selection to user-controlled input (the default \u2014 `i18next-browser-languagedetector` reads `?lng=` query params, cookies, `localStorage`, and request headers), an attacker can inject characters that change the structure of the outgoing request URL.\n\nAffected call sites:\n\n- `_readAny` \u2014 `lib/index.js:64`: `interpolate(resolvedLoadPath, { lng: languages.join(\u0027+\u0027), ns: namespaces.join(\u0027+\u0027) })`\n- `create` \u2014 `lib/index.js:123` (pre-patch): `interpolate(addPath, { lng, ns: namespace })`\n\nThe helper `interpolate` (`lib/utils.js`) previously returned the raw value with no encoding. In contrast, `addQueryString` already correctly uses `encodeURIComponent` for each query-string param \u2014 only the URL-path substitution was unprotected.\n\n### Impact\n\nAn attacker who can influence the resolved `lng` or `ns` value can alter the URL in several ways:\n\n- **Path traversal** \u2014 `lng = \u0027../../config\u0027` turns `/locales/{{lng}}/{{ns}}.json` into `/locales/../../config/translation.json`. On a misconfigured web server, this can cause the request to target a different resource than intended; in SSR pipelines that use `file://` or similar schemes for `loadPath`, it can read arbitrary files from the host filesystem.\n- **Query-string injection** \u2014 `lng = \u0027en?admin=true\u0027` turns `/locales/{{lng}}/{{ns}}.json` into `/locales/en?admin=true/translation.json`. Some server frameworks parse the query portion with higher priority than the path and branch on attacker-controlled flags.\n- **Fragment truncation** \u2014 `lng = \u0027en#anything\u0027` silently discards the rest of the path in browser fetches (client cannot see the final URL).\n- **URL-encoded bypasses** \u2014 `lng = \u0027en%2F..\u0027`, after server-side URL decoding, resolves to `en/..` \u2014 the attacker bypasses the absence of a literal `/` in their input.\n\nThe practical worst case is **SSRF** when `loadPath` is an internal or file-scheme URL, and **path-based authorisation bypass** against servers that segment access by URL prefix.\n\n### Also fixed in 3.0.5\n\n- **Per-instance `omitFetchOptions`.** A module-level boolean in `lib/request.js` was flipped to `true` the first time any backend instance hit a \"not implemented\" fetch error. Once flipped, **all** subsequent requests from **all** backend instances in the same module silently stripped every user-configured fetch option \u2014 including security-relevant `credentials`, `mode`, and `cache`. One misbehaving instance (for example during SSR hydration or in React Native) permanently removed these protections process-wide. 3.0.5 scopes the flag to the backend\u0027s `options` object (`options._omitFetchOptions`) so one instance\u0027s fallback cannot pollute siblings.\n- **Log forging via control characters in `lng`/`ns`.** Error callbacks embedded the raw `lng`/`ns`/URL in the message string. Crafted CR/LF values could inject fake log lines into file-backed log aggregators (CWE-117). 3.0.5 strips C0/C1 control chars before concatenation.\n- **Basic-auth credentials leaked into error callbacks.** If `loadPath` contained a `user:password@host` authority, the full URL (including the credentials) ended up in the error message strings returned to the caller. 3.0.5 redacts `user:password@` before logging.\n- **Prototype-pollution amplification via `for...in`.** `addQueryString` and the XHR `customHeaders` loop used `for...in` which walks the prototype chain. Polluted `Object.prototype` entries could leak into URL query parameters and request headers. 3.0.5 uses `Object.keys` and an explicit prototype-key guard.\n\n### Affected versions\n\nAll versions of `i18next-http-backend` prior to **3.0.5**.\n\n### Patch\n\nFixed in **3.0.5**. Summary of the hardening:\n\n1. New `utils.interpolateUrl` (used by `_readAny` and `create`) returns `null` if any substitution fails the URL-segment safety check (blocks `..`, `/`, `\\`, `?`, `#`, `%`, `@`, whitespace, control chars, prototype keys, and values \u003e 128 chars). Multi-language joins (`en+de`) are validated per-segment. The call sites now refuse to issue a request when the check fails and call back with a clear error.\n2. `omitFetchOptions` is stored per-instance on `options._omitFetchOptions`.\n3. Error-callback messages sanitise strings and redact URL credentials.\n4. `for...in` over untrusted objects replaced with `Object.keys` + prototype-key guard.\n\n### Workarounds\n\nNo workaround short of upgrading. If you cannot upgrade immediately, sanitise `lng` / `ns` yourself before they reach i18next (strip `..`, `/`, `\\`, `?`, `#`, `%`, whitespace, and control characters; cap the length).\n\n### Credits\n\nDiscovered via an internal security audit of the i18next ecosystem.",
  "id": "GHSA-q89c-q3h5-w34g",
  "modified": "2026-05-11T13:29:17Z",
  "published": "2026-04-22T17:41:24Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/i18next/i18next-http-backend/security/advisories/GHSA-q89c-q3h5-w34g"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-41691"
    },
    {
      "type": "WEB",
      "url": "https://github.com/i18next/i18next-http-backend/commit/4cee84f229c637b9c182366d3156f726d407a621"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/i18next/i18next-http-backend"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": " i18next-http-backend has Path Traversal \u0026 URL Injection via Unsanitised lng/ns"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…