Research
MCPThreatHive: First Automated Threat Intelligence Platform for MCP Ecosystems
Researchers release MCPThreatHive, an open-source platform that automates threat detection across Model Context Protocol-based agent systems — the first tool purpose-built for MCP security at scale. This arrives as MCP crosses 97 million installs and independent audits reveal 43% of MCP servers contain command injection vulnerabilities, 33% allow unrestricted network access, and 5% of open-source servers are already seeded with tool-poisoning attacks. Any builder running MCP servers in production should audit their deployments immediately.
Source
↳ Follow the thread