MCP Security Notification: Tool Poisoning Attacks
We have discovered a critical vulnerability in the Model Context Protocol (MCP) that allows for
invariantlabs.ai