A severe security flaw was identified in the AI-Bolit component of Imunify security products, creating significant concern among Linux server and web-hosting administrators. The issue allowed arbitrary code execution and possible root-level privilege escalation due to unsafe logic in AI-Bolit’s deobfuscation functions. Attackers could craft malicious files or database entries that caused the scanner to execute unintended PHP functions, potentially leading to full server compromise. The vulnerability affected older versions of Imunify360, ImunifyAV+, and ImunifyAV released before 32.7.4-1. The flaw was responsibly reported by security researcher Aleksejs Popovs, prompting Imunify to investigate and publish a fix on October 23, 2025. The patch introduced a strict allowlist of safe functions to ensure the deobfuscator could no longer be manipulated into executing attacker-supplied code. Technical analysis revealed that functions such as deobfuscateDeltaOrd and deobfuscateEvalHexFunc invoked Helpers::executeWrapper() without adequate filtering, enabling arbitrary PHP execution during file or database scans. Imunify confirmed that most servers were automatically updated by November 17, 2025, and no exploitation attempts have been detected. Users are strongly advised to upgrade the AI-Bolit package to version 32.7.4-1 or later, or apply the backported fix for CentOS 6. Until patching is complete, administrators should disable all file-scanning features or restrict them to trusted users. Imunify emphasizes keeping auto-updates enabled to ensure rapid protection and acknowledges Popovs for his responsible disclosure. This incident highlights the need for timely patching and continuous monitoring, even for tools designed to provide security.
Cisco has disclosed a high-severity vulnerability identified as CVE-2026-20188 affecting Cisco Crosswork Network Controller (CNC) and Network Services Orchestrator (NSO) platforms....
Palo Alto Networks has warned customers about a critical security flaw affecting its PAN-OS firewall platform. The vulnerability, identified as CVE-2026-0300, is linked to the User...
In early 2026, Rapid7 observed a sophisticated cyberattack attributed to the Iranian state-sponsored group MuddyWater, masquerading under the Chaos ransomware brand in what appears...