Source: thehackernews.com
OX Security researchers scanned 15,465 public MCP (Model Context Protocol) servers and discovered critical vulnerabilities within the ecosystem. MCP, introduced in 2024, was designed as a universal standard for connecting AI models, agents, and IDEs to tools and data. Despite the protocol's rapid adoption by thousands of developers and enterprises, significant security gaps were found. The research team traced critical vulnerabilities in Anthropic's MCP infrastructure. The findings highlight systemic security risks in AI agent tooling ecosystems. The article underscores the danger of rushing AI integration standards into production without adequate security vetting. The scale of exposed servers indicates widespread enterprise and developer exposure. This represents a significant supply chain and emerging technology risk for organizations leveraging AI agent workflows.
