Introduction
Account hijacking in enterprise single sign-on (SSO) environments can give attackers access to sensitive business applications and data. CVE-2025-40758, a recent vulnerability in the Siemens Mendix SAML module, creates that risk by letting attackers bypass signature validation in SAML authentication flows.
Siemens works in industrial automation and digitalization, and Mendix is its low-code application platform. Mendix is widely used in critical infrastructure sectors, so flaws in its authentication modules carry particular weight for large enterprises and public-sector organizations.
Technical information
CVE-2025-40758 is a flaw in how the Mendix SAML module enforces signature validation and binding checks. The module does not sufficiently verify the cryptographic signature on SAML assertions or responses. This improper verification, categorized as CWE-347, lets an unauthenticated remote attacker craft SAML messages that bypass signature validation. In vulnerable SSO configurations, the result can be full account hijacking with no user interaction.
The vulnerability sits in the core SAML authentication logic and affects all Mendix SAML modules prior to the fixed versions. Any network-accessible SAML endpoint configured for SSO is part of the attack surface, and the flaw is remotely exploitable. No public code snippets or PoC are available for this issue.
Affected systems and versions
- Mendix SAML (Mendix 9.24 compatible): All versions before V3.6.21
- Mendix SAML (Mendix 10.12 compatible): All versions before V4.0.3
- Mendix SAML (Mendix 10.21 compatible): All versions before V4.1.2
Any SSO setup that uses one of the affected SAML modules is a vulnerable configuration.
Vendor security history
Siemens Mendix SAML modules have had similar authentication and signature validation vulnerabilities before. CVE-2023-25957, for example, involved insufficient assertion verification and received a high CVSS score. Siemens typically responds with advisories and patches through its ProductCERT process. The recurrence of SAML-related flaws, though, suggests ongoing architectural and implementation weaknesses in this part of the platform.
