The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has updated its Known Exploited Vulnerabilities (KEV) catalog, adding seven critical security weaknesses that have recently come under attack. This update was announced on Wednesday, highlighting the growing threat landscape in cybersecurity.
The vulnerabilities include serious threats such as a server-side request forgery in SonicWall SMA 1000 Appliances and a command injection flaw in Kestra OSS, both with the highest CVSS score of 10.0. These weaknesses could potentially allow unauthorized access and remote code execution by attackers.
SonicWall and Sangoma Vulnerabilities Under Fire
Among the vulnerabilities, CVE-2026-83548 and CVE-2026-83549 in SonicWall SMA 1000 Appliances have drawn attention due to their severe impact. These flaws allow unauthorized operations and arbitrary command execution, leading to significant security breaches.
Similarly, a critical SQL injection vulnerability, CVE-2026-9586, in Sangoma Switchvox could enable attackers to manipulate backend databases, potentially resulting in remote code execution and further exploitation of systems.
Threats to JFrog Artifactory and Kestra OSS
The JFrog Artifactory is also vulnerable to improper authentication, as identified in CVE-2026-82329. This flaw could grant attackers administrative privileges, exacerbating the potential damage from unauthorized network access.
In the case of Kestra OSS, CVE-2026-49869 allows attackers to execute arbitrary workflows, leading to extensive breaches including cryptocurrency mining and data theft. These vulnerabilities are of particular concern due to the potential impact on critical infrastructure.
AI Infrastructure and Persistent Threats
The highlighted vulnerabilities reveal a trend towards targeting AI infrastructure. For example, flaws in LiteLLM are being exploited to deploy cryptocurrency miners and acquire sensitive data. The ongoing attacks emphasize the necessity for robust security measures in AI systems.
Federal agencies have been advised to prioritize patching these vulnerabilities, especially given the aggressive tactics of threat actors. The deadline for updates is set for September 16, 2026, for most vulnerabilities, highlighting the urgency of the situation.
As cyber threats become increasingly sophisticated, organizations must stay vigilant and proactive in their security efforts. Monitoring and updating AI workloads and backend systems are crucial steps in safeguarding against these persistent threats.
