#軟體漏洞

1 articles

A conceptual, dark-themed image of a futuristic server farm where glowing data lines are being inter

Researchers Identify Vulnerability in Anthropic's Model Context Protocol

Researchers have uncovered a critical command execution flaw in Anthropic's Model Context Protocol (MCP) that exposes nearly 200,000 servers, highlighting urgent security risks for AI agent-tool communication.

Jason·May 3, 2026·2 min read