GHSA-q89c-q3h5-w34g
ADVISORY - githubSummary
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:
_readAny—lib/index.js:64:interpolate(resolvedLoadPath, { lng: languages.join('+'), ns: namespaces.join('+') })create—lib/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 traversal —
lng = '../../config'turns/locales/{{lng}}/{{ns}}.jsoninto/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 usefile://or similar schemes forloadPath, it can read arbitrary files from the host filesystem. - Query-string injection —
lng = 'en?admin=true'turns/locales/{{lng}}/{{ns}}.jsoninto/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 truncation —
lng = 'en#anything'silently discards the rest of the path in browser fetches (client cannot see the final URL). - URL-encoded bypasses —
lng = 'en%2F..', after server-side URL decoding, resolves toen/..— 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 inlib/request.jswas flipped totruethe 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-relevantcredentials,mode, andcache. 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'soptionsobject (options._omitFetchOptions) so one instance's fallback cannot pollute siblings. - Log forging via control characters in
lng/ns. Error callbacks embedded the rawlng/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
loadPathcontained auser:password@hostauthority, the full URL (including the credentials) ended up in the error message strings returned to the caller. 3.0.5 redactsuser:password@before logging. - Prototype-pollution amplification via
for...in.addQueryStringand the XHRcustomHeadersloop usedfor...inwhich walks the prototype chain. PollutedObject.prototypeentries could leak into URL query parameters and request headers. 3.0.5 usesObject.keysand 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:
- New
utils.interpolateUrl(used by_readAnyandcreate) returnsnullif 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. omitFetchOptionsis stored per-instance onoptions._omitFetchOptions.- Error-callback messages sanitise strings and redact URL credentials.
for...inover untrusted objects replaced withObject.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.
GitHub
CVSS SCORE
6.5medium| Package | Type | OS Name | OS Version | Affected Ranges | Fix Versions |
|---|---|---|---|---|---|
| i18next-http-backend | npm | - | - | <3.0.5 | 3.0.5 |
CVSS:3 Severity and metrics
The CVSS metrics represent different qualitative aspects of a vulnerability that impact the overall score, as defined by the CVSS Specification.
The vulnerable component is bound to the network stack, but the attack is limited at the protocol level to a logically adjacent topology. This can mean an attack must be launched from the same shared physical (e.g., Bluetooth or IEEE 802.11) or logical (e.g., local IP subnet) network, or from within a secure or otherwise limited administrative domain (e.g., MPLS, secure VPN to an administrative network zone). One example of an Adjacent attack would be an ARP (IPv4) or neighbor discovery (IPv6) flood leading to a denial of service on the local LAN segment (e.g., CVE-2013-6014).
Specialized access conditions or extenuating circumstances do not exist. An attacker can expect repeatable success when attacking the vulnerable component.
The attacker is unauthorized prior to attack, and therefore does not require any access to settings or files of the vulnerable system to carry out an attack.
The vulnerable system can be exploited without interaction from any user.
An exploited vulnerability can only affect resources managed by the same security authority. In this case, the vulnerable component and the impacted component are either the same, or both are managed by the same security authority.
There is some loss of confidentiality. Access to some restricted information is obtained, but the attacker does not have control over what information is obtained, or the amount or kind of loss is limited. The information disclosure does not cause a direct, serious loss to the impacted component.
Modification of data is possible, but the attacker does not have control over the consequence of a modification, or the amount of modification is limited. The data modification does not have a direct, serious impact on the impacted component.
There is no impact to availability within the impacted component.