Latest MCP News (www.gentoro.com)

The latest MCP Weekly Digest reveals significant advancements in AI and ML as Anthropic contributed its MCP technology to the Agentic AI Foundation (AAIF) under the Linux Foundation, while also launching a $320 million partnership with the U.S. Department of Energy called the "Genesis Mission." This initiative aims to enhance American research productivity by employing MCP and new "Skills" that encapsulate institutional knowledge as reusable instruction sets. Concurrently, a critical security vulnerability (CVE-2025-64106) was identified in the Cursor IDE, underscoring the urgent need to prioritize security as AI installations become more commonplace. Anthropic’s separation of MCP from the new Agent Skills standard marks a pivotal shift in how agents interact with tools, enhancing efficiency and clarity in complex tasks. OpenAI's GPT-5.2-Codex has been optimized for agentic software engineering, bringing sophisticated abilities to long-term project management through "context compaction." In response to rising security threats, OpenAI also updated ChatGPT Atlas with advanced defenses against prompt injection, utilizing adversarial training and reinforcement learning. With the increasing integration of AI into critical sectors like national research, ensuring robust security protocols and maintaining user trust are now essential to the future of agentic AI technologies.