Summary
Froxlor customer can create MySQL databases on disallowed servers via Mysqls.add API
Full technical description
The Mysqls.add API command (lib/Froxlor/Api/Commands/Mysqls.php) accepts a customer-controlled mysql_server parameter and only validates that the value is numeric and that the server index exists in userdata.inc.php. It never checks the value against the calling customer's allowed_mysqlserver allowlist. A customer can therefore create a database, plus a MySQL user with a password they choose, on any MySQL server the operator has configured, including servers that were explicitly excluded from that customer (e.g. a separate cluster, premium-tier host, or another tenant pool). The same allowed_mysqlserver check is correctly enforced in MysqlServer::get() / MysqlServer::listing() and in the customer-facing UI (customer_mysql.php), confirming the omission is a bug, not by-design.
Details
Vulnerable code path, lib/Froxlor/Api/Commands/Mysqls.php:69-99 (add()):
public function add()
{
if (($this->getUserDetail('mysqls_used') < $this->getUserDetail('mysqls') || ...) {
...
$customer = $this->getCustomerData('mysqls'); // line 80
$dbserver = $this->getParam('mysql_server', true, // line 81, user-controlled
$this->getDefaultMySqlServer($customer));
...
$dbserver = Validate::validate($dbserver, ..., '/^[0-9]+$/', ...); // line 92, numeric only
Database::needRoot(true, $dbserver, false); // line 93, root ctx for ANY index
Database::needSqlData();
$sql_root = Database::getSqlData();
Database::needRoot(false);
if (!is_array($sql_root)) { // line 97, only existence check
throw new Exception("Database server with index #" . $dbserver . " is unknown", 404);
}
...
$username = $dbm->createDatabase($newdb_params['loginname'], $password,
$dbserver, ...); // line 116/118, DB+user created
...
Database::pexecute($stmt, ["customerid"=>$customer['customerid'], ..., "dbserver"=>$dbserver], ...);
}
}
The $customer['allowed_mysqlserver'] field IS read on line 80 but is only consumed by getDefaultMySqlServer() (lines 566-573) to compute a default when the request omits mysql_server. As soon as the client supplies the parameter, the default path is skipped and no further authorization gate runs.
Cross-file evidence the check is intended elsewhere:
lib/Froxlor/Api/Commands/MysqlServer.php:319-323,get()rejects with HTTP 405 when$dbserveris not inallowed_mysqlserver:if ($this->isAdmin() == false) { $allowed_mysqls = json_decode($this->getUserDetail('allowed_mysqlserver'), true); if ($allowed_mysqls === false || empty($allowed_mysqls) || !in_array($dbserver, $allowed_mysqls)) { throw new Exception("You cannot access this resource", 405); } ... }lib/Froxlor/Api/Commands/MysqlServer.php:252-257, same allowlist filter onlisting().customer_mysql.php:222, UI rejects withResponse::dynamicError('No permission')whenempty($allowed_mysqlservers).
Chain of execution (attacker → impact):
- Customer authenticates to
api.phpwith apikey/secret. The only API gate iscust_api_allowed;allowed_mysqlserveris not consulted at auth time. - Customer sends JSON
{"command":"Mysqls.add","params":{"mysql_password":"<valid>","mysql_server":<disallowed_idx>}}. Mysqls.php:71quota check passes (mysqls_used < mysqls).Mysqls.php:80getCustomerData('mysqls')returns the caller's own row.Mysqls.php:81$dbserveris set from the request (default-fallback path skipped).Mysqls.php:92numeric regex passes.Mysqls.php:93-99Database::needRoot(true, $dbserver, false)switches to the root context of the attacker-chosen server; existence check passes.Mysqls.php:116/118DbManager::createDatabase(...)runs against the disallowed server using stored root credentials, creating the DB and granting the supplied password to<loginname>_<sqlN>(DbManager.php:177-218).Mysqls.php:127-141inserts a row intoTABLE_PANEL_DATABASESwith the attacker'scustomeridand the disalloweddbserver, allowing later management viaMysqls.get/update/delete(which only filter bycustomeridfor non-admins, e.g.Mysqls.php:282).
PoC
Preconditions on the target instance:
- ≥2 MySQL servers configured in
lib/userdata.inc.php(e.g. index 0 default, index 1 internal/premium). - Customer X with
allowed_mysqlserver=[0],cust_api_allowed=1,mysqls > 0, and an issued API key (apikey:secret).
Request, customer creates a database on server 1, which is not in their allowlist:
curl -k -u 'CUST_APIKEY:CUST_SECRET' \
-H 'Content-Type: application/json' \
-X POST \
-d '{"command":"Mysqls.add","params":{"mysql_password":"ValidP@ssw0rd!","mysql_server":1}}' \
https://froxlor.example.com/api.php
Expected (mirroring MysqlServer.get() behaviour): HTTP 405, "You cannot access this resource".
Actual: HTTP 200 with the full database record, e.g.:
{"data":{"id":42,"customerid":<cust_id>,"databasename":"<loginname>_sql1","dbserver":1,...}}
Verify the credentials work on the forbidden server:
mysql -h server1.host -u <loginname>_sql1 -p # password: ValidP@ssw0rd!
mysql> SHOW DATABASES; # the new DB is present
mysql> USE <loginname>_sql1; # full access to the newly-created DB
The customer can subsequently manage the DB via Mysqls.get, Mysqls.update, and Mysqls.delete, those non-admin code paths filter only by customerid (Mysqls.php:282-289, Mysqls.php:380-391), which matches.
Impact
- Bypass of the per-customer MySQL-server allowlist (
allowed_mysqlserver) enforced by the admin/reseller. The authorization model is fully defeated for theaddoperation. - The customer obtains valid MySQL credentials on a server the operator explicitly excluded for them, possibly an internal/separate cluster, billing tier, premium-only host, or a server provisioned for a different tenant pool.
- The customer can persist a DB on the forbidden server (resource and policy bypass), then read/write data there, and continue to manage it through
Mysqls.update/Mysqls.delete. - Impact is bounded: privileges granted by
DbManager::grantPrivilegesToapply only to the new<loginname>_sqlNdatabase, so no cross-tenant data exposure on the forbidden server. The damage is policy bypass, resource consumption on the forbidden server, and credential persistence there.
GHSA-Q4RM-M6XH-5PV7 has a CVSS score of 4.3 (Medium). The vector is network-reachable, low privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (2.3.7); upgrading removes the vulnerable code path.
Affected versions
Security releases
Kodem intelligence
Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.
Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.
Already deployed Kodem?
See it in your environmentNew to Kodem? Get a demo →Remediation advice
Mirror the allowlist check already present in MysqlServer::get(). After the numeric validation on Mysqls.php:92, before Database::needRoot(...), add for non-admin callers:
// validate whether the dbserver exists
$dbserver = Validate::validate($dbserver, html_entity_decode(lng('mysql.mysql_server')), '/^[0-9]+$/', '', 0, true);
// enforce per-customer allowed_mysqlserver allowlist (parity with MysqlServer::get())
if (!$this->isAdmin()) {
$allowed = json_decode($customer['allowed_mysqlserver'] ?? '[]', true);
if (!is_array($allowed) || empty($allowed)
|| !in_array((int)$dbserver, array_map('intval', $allowed), true)) {
throw new Exception('You cannot access this resource', 405);
}
}
Database::needRoot(true, $dbserver, false);
Audit Mysqls::update(), Mysqls::delete(), and Mysqls::get() for the same gap: those endpoints accept mysql_server and ultimately call Database::needRoot(true, $result['dbserver'], false) on the row's stored value. Once the row exists with a forbidden dbserver, those paths execute against the forbidden server unchallenged. Consider rejecting any non-admin operation whose target row's dbserver is outside allowed_mysqlserver, even if the row already exists, to defend in depth.
Frequently Asked Questions
- What is GHSA-Q4RM-M6XH-5PV7? GHSA-Q4RM-M6XH-5PV7 is a medium-severity security vulnerability in froxlor/froxlor (composer), affecting versions <= 2.3.6. It is fixed in 2.3.7.
- How severe is GHSA-Q4RM-M6XH-5PV7? GHSA-Q4RM-M6XH-5PV7 has a CVSS score of 4.3 (Medium). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
- Which versions of froxlor/froxlor are affected by GHSA-Q4RM-M6XH-5PV7? froxlor/froxlor (composer) versions <= 2.3.6 is affected.
- Is there a fix for GHSA-Q4RM-M6XH-5PV7? Yes. GHSA-Q4RM-M6XH-5PV7 is fixed in 2.3.7. Upgrade to this version or later.
- Is GHSA-Q4RM-M6XH-5PV7 exploitable, and should I be worried? Whether GHSA-Q4RM-M6XH-5PV7 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
- What actually determines whether GHSA-Q4RM-M6XH-5PV7 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
- How do I fix GHSA-Q4RM-M6XH-5PV7? Upgrade
froxlor/froxlorto 2.3.7 or later.