Learn how to secure Model Context Protocol proxies with post-quantum cryptographic agility. Protect AI infrastructure against future quantum threats with hybrid encryption.

Anthropic sees no issues - and says the tools are working as intended.

安全研究团队OX Security近日披露，由Anthropic公司开发并维护的AI通信协议标准MCP（Model Context Protocol）存在重大安全隐患。该协议的架构设计缺陷可能导致服务器遭受远程代码执行（RCE）攻击，威胁范围覆盖Python、Typescript、Java和Rust等所有支持语言，所有基于MCP构建的AI系统均面临潜在风险。 研究团队通过技术分析发现，漏洞源于协议 ...

A design flaw – or expected behavior based on a bad design choice, depending on who is telling the story – baked into ...

IT之家 4 月 18 日消息，安全研究团队 OX Security 本周（4 月 15 日）发现，Anthropic 创建、维护的行业标准 AI 通信协议 MCP（IT之家注：Model Context ...

However, in a report published on April 15, researchers at Ox Security claimed that a flaw in the protocol could enable ...

Unsafe defaults in MCP configurations open servers to possible remote code execution, according to security researchers who ...

While Anthropic's dispute with the Pentagon escalated over guardrails on military use, OpenAI LLC struck its own publicized ...

Solution also includes a new COM/Python API that exposes the simulation engine to external automation, the company explains..

AI开发社区近日神经紧绷，一份安全报告指出，超过20万台AI服务器面临潜在攻击风险。 网络安全公司OX Security于2026年4月15日发布报告，揭示了Anthropic的MCP（Model Context Protocol，模型上下文协议）存在设计缺陷，可能导致远程代码执行。该漏洞影响范围广泛，波及超过20万台AI服务器。 MCP是由Anthropic于2024年11月推出的开源标准，旨在 ...

IT之家 4 月 16 日消息，网络安全公司 OX Security 昨日（4 月 15 日）发布报告，披露 Anthropic 的 MCP（模型上下文协议）存在设计缺陷，可导致远程代码执行。该设计缺陷影响范围极广，导致超过 20 万台 AI ...