Cybersecurity researchers have uncovered the first known malicious Model Context Protocol (MCP) server in the wild. The incident involved a compromised npm package, “postmark-mcp,” which appeared legitimate but contained malicious code. The code revealed significant supply chain security vulnerabilities. Version 1.0.16 released in September 2025, secretly forwarded emails to an attacker-controlled address, which exposed sensitive data. This incident underscores how open-source components can be exploited within enterprise environments. Especially where MCP servers manage critical information like passwords and customer data. Simultaneously, threat actors launched a highly sophisticated phishing campaign impersonating Ukrainian government agencies. They used malicious SVG files embedded in emails falsely claiming to be police notices. These files redirect victims to download password-protected ZIP archives. The files contained CHM files that activate CountLoader, which is a versatile malware loader capable of deploying tools like Cobalt Strike, Amatera Stealer, and PureMiner. Employing advanced techniques such as process hollowing and in-memory execution, these malware strains evade traditional defenses. They appear to have been targeting Ukrainian entities with layered, persistent attack methods. CTIX analysts will continue to report on the latest malware strains and attack methodologies.