GHSA-8847-338w-5hcj
ADVISORY - githubSummary
Summary
Versions of i18next-fs-backend prior to 2.6.4 interpolate the caller-supplied lng and ns values directly into the configured loadPath and addPath templates with no path-component validation and no sanitisation. When an application exposes the resolved language code to user-controlled input (?lng= query parameter, cookie, request header), a crafted value can break out of the intended locale directory.
Affected call sites in lib/index.js:
read(line 38 pre-patch):const filename = interpolate(loadPath, { lng: language, ns: namespace })removeFile(line 101 pre-patch): same pattern againstaddPathwriteFile(line 127 pre-patch): same pattern againstaddPathfor queued missing-key writes
The helper interpolate in lib/utils.js substitutes raw values with no encoding — unlike the addQueryString helper in i18next-http-backend, there is no equivalent safety for path interpolation.
Impact
- Arbitrary file read. With a
loadPathlike/locales/{{lng}}/{{ns}}.json, an attacker-controlledlng = '../../etc'(and matchingns) causes the backend to read a file outside the locale directory. For parsers that tolerate arbitrary content (YAML's freeform text), the file contents surface as a translation resource. - Arbitrary file overwrite.
addPathis interpolated the same way for missing-key writes (thecreate()code path and the debounced writer inwriteFile). A traversinglng/nscombination can cause the process to write JSON structures to an unintended filesystem location, potentially overwriting application files if the process user has write access. - Chain with
.js/.tseval.i18next-fs-backendsupports loading.jsand.tslocale files byeval-ing their content (intentional feature, documented as requiring trusted sources). Combining traversal with that path — for examplelng = '../../../app/config'againstloadPath: '/locales/{{lng}}/{{ns}}.js'— would cause the backend to execute a server-side file as JavaScript, exfiltrating whatever it can touch (process.env, connected services).
Exploitation requires the application to pass an untrusted lng/ns value through to i18next.t() without its own validation. Many i18next setups do exactly this via i18next-browser-languagedetector (query string / cookie detection).
Affected versions
All versions of i18next-fs-backend prior to 2.6.4.
Patch
Fixed in 2.6.4. lib/utils.js now exports:
isSafePathSegment(v)— returnstrueonly ifvis a non-empty string of ≤ 128 chars that does not contain..,/,\, control characters, or a prototype key (__proto__,constructor,prototype). Legitimate i18next language-code shapes (BCP-47,en_US,zh-Hant-HK,pirate-speak,my-custom.ns,+-joined multi-language values) all pass.interpolatePath(template, data)— substitutes variables like the existinginterpolatebut refuses the whole result if any segment failsisSafePathSegment. Callers bail out with an error (read) or silently drop the queued write (writeFile,removeFile).
The .js / .ts eval behaviour is intentionally retained — dynamic expressions in locale files are a documented feature of this backend, and safe replacements like dynamic import() are async-only and incompatible with this backend's sync-capable code path. The README has a new "Security considerations" section that spells out the trust model: .js/.ts locale files must be treated as code.
Workarounds
No workaround short of upgrading. If you cannot upgrade immediately, sanitise lng / ns at your application boundary before passing them to i18next — reject values containing .., /, \, control characters, and cap the length.
Credits
Discovered via an internal security audit of the i18next ecosystem.
GitHub
CVSS SCORE
8.2high| Package | Type | OS Name | OS Version | Affected Ranges | Fix Versions |
|---|---|---|---|---|---|
| i18next-fs-backend | npm | - | - | <2.6.4 | 2.6.4 |
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 a total loss of confidentiality, resulting in all resources within the impacted component being divulged to the attacker. Alternatively, access to only some restricted information is obtained, but the disclosed information presents a direct, serious impact. For example, an attacker steals the administrator's password, or private encryption keys of a web server.
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.