Description

ModSecurity, one of the most widely deployed open-source web application firewalls (WAFs), has received security updates addressing two vulnerabilities that could allow attackers to evade security rules using carefully constructed HTTP requests. Tracked as CVE-2026-52761 and CVE-2026-52747, the flaws impact ModSecurity releases through version 3.0.15 and have been resolved in version 3.0.16. Both issues stem from inconsistencies in request processing and input transformation, creating situations where malicious traffic may be interpreted differently by the WAF and the protected application. Such discrepancies can reduce the effectiveness of security policies and increase the risk of malicious requests reaching backend systems without detection. The first vulnerability, CVE-2026-52761, affects the `t:utf8toUnicode` transformation used by ModSecurity rule sets. The problem originates from the way memory size is determined during Unicode conversion, resulting in inconsistent output on different processor architectures. On 32-bit i386 systems, the transformation can generate incomplete or inaccurate Unicode values, while 64-bit platforms are not affected because the conversion behaves as expected in those environments. Consequently, detection rules that depend on this transformation may fail to identify specially crafted payloads, creating an opportunity for attackers to bypass inspection. The flaw was discovered by the Core Rule Set (CRS) team during automated testing. Although there are no known reports of active exploitation, the vulnerability can be triggered remotely without authentication, making it important for organizations to deploy the available fixes or transition away from affected 32-bit systems. The second flaw, CVE-2026-52747, is considered more serious because it affects ModSecurity's multipart/form-data parser. During request inspection, embedded carriage return and line feed characters within non-file form fields are removed before values are exposed to WAF rules, while backend applications may still process the original input. This parsing mismatch enables attackers to conceal payloads that depend on line breaks, potentially bypassing detection signatures designed to identify multiline injection attempts. Researchers also found that ModSecurity's strict multipart validation does not flag these malformed requests, leaving the bypass unnoticed. Organizations are advised to upgrade to ModSecurity 3.0.16, review existing WAF rules, strengthen server-side input validation, and perform regression testing to ensure request parsing remains consistent across both the WAF and backend applications.