Description

Three recently revealed vulnerabilities in the Sitecore Experience Platform (SXP) are critical security threats, with the possibility of information disclosure and remote code execution (RCE). The flaws were discovered by watchTowr Labs and are listed as CVE-2025-53693 (HTML cache poisoning through unsafe reflections), CVE-2025-53691 (RCE through insecure deserialization), and CVE-2025-53694 (information disclosure within the Item Service API enabling cache key enumeration). Whereas patches for the first two were rolled out in June 2025, the third was resolved in July. Sitecore has confirmed that successful exploitation of the vulnerabilities can enable attackers to run arbitrary code or access otherwise unauthorized sensitive data. These fresh vulnerabilities are based on previous flaws similarly detected by watchTowr, such as CVE-2025-34509 (use of hard-coded credentials), CVE-2025-34510 (RCE through path traversal), and CVE-2025-34511 (RCE through Sitecore PowerShell Extension). With CVSS ratings of 8.2 and 8.8, these disclosed vulnerabilities already showed the platform vulnerable to attack, especially in post-authentication scenarios. Introducing the pre-authentication cache poisoning vulnerability further magnifies the threat level by enabling attackers to launch an exploit chain even without prior access. A threat actor could, according to researcher Piotr Bazydlo, brute-force HTML cache keys exposed by the ItemService API and then poison these keys with malicious payloads. In conjunction with CVE-2025-53691, the attacker can inject specially crafted HTML to deserialization through the .NET BinaryFormatter, resulting in full code execution. This attack allows the hijacking of Sitecore pages and the injection of arbitrary JavaScript to post-auth RCE vulnerabilities. The study emphasizes the immediate need for the administrators to implement all the available patches and audit external API exposure to minimize the attack surface.