@vendure/core has a SQL Injection vulnerability

Description

Summary

An unauthenticated SQL injection vulnerability exists in the Vendure Shop API. A user-controlled query string parameter is interpolated directly into a raw SQL expression without parameterization or validation, allowing an attacker to execute arbitrary SQL against the database. This affects all supported database backends (PostgreSQL, MySQL/MariaDB, SQLite).

The Admin API is also affected, though exploitation there requires authentication.

Affected versions

  • @vendure/core < 2.3.4
  • @vendure/core >= 3.0.0, < 3.5.7
  • @vendure/core >= 3.6.0, < 3.6.2

Note: versions 2.3.4 and above in the 2.x line are patched. There were no 2.4.x or 2.x releases between 2.3.x and 3.0.0.

Patched versions

  • @vendure/core 2.3.4
  • @vendure/core 3.5.7
  • @vendure/core 3.6.2

Details

In ProductService.findOneBySlug, the request context's languageCode value is interpolated into a SQL CASE expression via a JavaScript template literal:

.addSelect(
    `CASE translation.languageCode WHEN &#x27;${ctx.languageCode}&#x27; THEN 2 WHEN &#x27;${ctx.channel.defaultLanguageCode}&#x27; THEN 1 ELSE 0 END`,
    &#x27;sort_order&#x27;,
)

TypeORM has no opportunity to parameterize this value because it is embedded directly into the SQL string before being passed to the query builder.

The languageCode value can originate from the HTTP query string and is set on the request context for every incoming API request. The value is cast to the LanguageCode TypeScript type at compile time, but no runtime validation is performed -- the raw query string value is used as-is.

Attack vector

An unauthenticated attacker can append a crafted languageCode query parameter to any Shop API request to inject arbitrary SQL into the query. No user interaction is required. The vulnerable endpoint is exposed on every default Vendure installation.

Mitigation

Upgrade to a patched version immediately.

If you cannot upgrade right away, apply the following hotfix to RequestContextService.getLanguageCode to validate the languageCode input at the boundary. This blocks injection payloads before they can reach any query:

private getLanguageCode(req: Request, channel: Channel): LanguageCode | undefined {
    const queryLanguageCode = req.query?.languageCode as string | undefined;
    const isValidFormat = queryLanguageCode &amp;&amp; /^[a-zA-Z0-9_-]+$/.test(queryLanguageCode);
    return (
        (isValidFormat ? (queryLanguageCode as LanguageCode) : undefined) ??
        channel.defaultLanguageCode ??
        this.configService.defaultLanguageCode
    );
}

This replaces the existing getLanguageCode method in packages/core/src/service/helpers/request-context/request-context.service.ts. Invalid values are silently dropped and the channel's default language is used instead.

The patched versions additionally convert the vulnerable SQL interpolation to a parameterized query as defense in depth.

Basic information

Type
reviewed
Severity
critical
Advisory on GitHub
Open advisory ↗
Repository advisory
Open repository advisory ↗
Source code
Browse source ↗
Published (advisory)
2026-04-14 22:38:01 UTC
Updated
2026-04-24 21:10:19 UTC
GitHub reviewed
2026-04-14 22:38:01 UTC
NVD published
2026-04-21 20:17:02 UTC

EPSS Score

Score Percentile
4.56% 89.22%

CVSS Scores

Base score Version Severity Vector
9.1 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H Click to expand
Attack vector (AV:N)
Could be attacked over the internet or any normal routed network—not just someone sitting at the machine.
Attack complexity (AC:L)
Once they can reach the bug, pulling it off is straightforward—no weird race conditions or rare setup.
Privileges required (PR:N)
No account or special rights needed—anonymous or random user is enough.
User interaction (UI:N)
Nobody has to click “OK” or open a trap file; it can work without a victim helping.
Scope (S:U)
Damage stays in the same “trust bubble” as the broken component—no big spill into unrelated systems.
Confidentiality (C:H)
Serious risk that confidential data gets exposed in a big way.
Integrity (I:N)
Data isn’t meaningfully altered or forged.
Availability (A:H)
Could take the service down hard or make it unusable for people who depend on it.

Identifiers

CWEs

CWE id Name
CWE-89 Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection')

Credits

  • jacobfrantz1 (reporter)

Affected packages (3)

Vulnerable version ranges and first patched releases as published by GitHub.

Ecosystem Package Vulnerable range First patched Vulnerable functions
npm @vendure/core >= 3.0.0, < 3.5.7 3.5.7
npm @vendure/core >= 3.6.0, < 3.6.2 3.6.2
npm @vendure/core >= 1.7.4, < 2.3.4 2.3.4

References

cvelogic Threat Intelligence