A newly revealed vulnerability in Fortinet's FortiWeb web application firewall, identified as CVE-2025-52970, poses a serious security risk. Discovered by researcher Aviv Y and named "FortMajeure," the flaw arises from an out-of-bounds read in the firewall's cookie parsing logic. By modifying the Era parameter, an attacker can force the server to fall back on an all-zero secret key used for session encryption and HMAC verification. This enables the creation of valid, forged authentication cookies, allowing full impersonation of any active user, including administrators. To exploit this flaw, an attacker needs the target user to be actively logged in. The attacker is required to identify a short numeric value embedded in the session cookie, which is validated through the `refresh_total_logins()` function. Although classified as a high-complexity attack, the guessing range is typically under 30, making brute-force attempts fast and efficient. Exploiting the all-zero encryption key allows attackers to instantly validate each attempt, greatly accelerating the overall attack workflow. The security flaw impacts FortiWeb versions ranging from 7.0 up to 7.6. Versions starting from 8.0 are not impacted. No temporary fixes or workarounds have been offered, and upgrading remains the only effective countermeasure. Although the CVSS rating stands at 7.7, the simplicity of the exploitation in practice makes the vulnerability more severe than the score indicates. Aviv Y released a partial proof of concept to demonstrate the issue but has withheld full exploit details to allow administrators time to secure their systems. He noted that creating a complete exploit requires in-depth reverse engineering of Fortinet’s proprietary session structure. However, the urgency to patch remains high, as cyber attackers often act quickly after public disclosures.
APT28 (also known as Fancy Bear), a Russia-linked advanced persistent threat group, has launched a targeted cyber espionage campaign leveraging a newly identified malware framework...
A large-scale data theft campaign has impacted more than a dozen companies following a breach at a SaaS integration provider, leading to the compromise of authentication tokens. Th...
hropic has introduced Claude Mythos Preview, an advanced general-purpose language model that demonstrates a striking ability to independently locate and exploit previously unknown ...