AI Development Tools Targeted by Search Poisoning and Malware Campaigns; MCP Vulnerabilities Expose Millions

Threat actors have turned the digital infrastructure of AI development—search results, package repositories, IDE integrations—into an attack surface, combining search-engine manipulation, supply-chain compromise, and protocol-level vulnerabilities to target developers at scale. Over the first half of 2026, infostealer malware campaigns masquerading as legitimate AI coding assistants reached more than 3.3 million protected users, while simultaneous vulnerabilities in Model Context Protocol implementations expose developers using Cursor, GitHub Copilot, Claude Code, and Amazon Q to arbitrary code execution without user interaction.