It turns out that the Model Context Protocol (MCP), a protocol used to connect generative AI models to other tools, has a vulnerability that could lead to the leakage of sensitive information, such as ...