Critical RCE Wave Hits llama.cpp and MCP Frameworks

A surge of critical vulnerabilities targets core AI infrastructure, including sandbox escapes in Model Context Protocol and multiple memory corruption flaws in llama.cpp enabling remote code execution.