OpenAI co-founds Agentic AI Foundation, donates AGENTS.md

今天， OpenAI 与 Anthropic 和 Block 在 Linux Foundation 之下共同发起了 Agentic AI Foundation ( AAIF )，并得到 Google、 Microsoft、 AWS、 Bloomberg 和 Cloudflare 的支持。 AAIF 的目标是在具有代理能力的 AI 系统从试验阶段走向实际部署时，为开放且可互操作的基础设施提供中立的管理与维护。

作为这一努力的一部分， OpenAI 将把 AGENTS.md 捐赠给该基金会。 AGENTS.md 是一种简单的开放格式文件，用来向智能代理提供项目特定的指令与上下文，捐赠旨在确保这一格式能在社区中长期得到支持与采纳。

为什么开放标准重要

开发者正在快速采用 AI 来构建更有能力的代理系统——从代码助手到工作流自动化和客服代理。到 2025 年，这些系统已开始从原型转为在企业与消费者场景中承担实际任务。

我们认为，要让这些代理从实验性工具平稳、大规模地过渡为现实系统，开放标准至关重要。开放标准有助于增强代理的安全性、降低开发难度、提高在不同工具与平台之间的可移植性，同时能防止随着这一新类别成熟而导致的生态分裂。

随着越来越多的代理开始承担真实职责，分裂造成的代价也随之上升。没有共同约定与中立治理，代理开发可能会分裂成互不兼容的孤岛，限制可移植性、安全性以及整体进步。 AAIF 的成立正是为了解决这一关键时刻出现的问题。

构建开放生态

过去一年， OpenAI 在开放构建代理基础设施方面做出了重要贡献——与开发者、研究者和其他公司协作，证明开放协议在实践中是可行的。

我们贡献了若干核心构件，包括 Agents SDK、 Apps SDK 和 Agentic Commerce Protocol，以及开源项目如 gpt-oss 和 Codex CLI。仅 Codex CLI 就已被用于帮助合并超过两百万个公开的 GitHub pull request。

我们也早期采用并作为核心贡献者参与了 Model Context Protocol ( MCP )，并将其作为 ChatGPT 中连接器和应用的基础。就在上周，我们与 Anthropic 和 MCP-UI 合作，通过 MCP Apps 将 Apps SDK 扩展到所有 MCP 开发者，使得在不同平台上构建互动型、由代理驱动的体验变得更容易。

这些努力证明了开放、可互操作的基础设施能够在工具和社区中获得实际采用，也为 AAIF 打下了基础。

捐赠 AGENTS.md

AGENTS.md 最初是为了解决一个常见的实际问题： Codex 需要一种可预测的方式来找到项目特定的指令——比如编码规范、构建步骤和测试要求——以便代理能在代码库中安全且高效地工作。

我们引入的 AGENTS.md 是一种轻量的 Markdown 文件，设计为与 README.md 并置，提供一种简单且可互操作的标准，给 AI 编码代理一个一致的项目级指导来源，使其在不同仓库和构建链中都能可靠运行。这种基于 Markdown 的约定极大提升了代理在多样化仓库和构建系统中的行为可预测性。

从内部方案起步后， AGENTS.md 很快在生态系统中得到认可。自 2025 年 8 月发布以来， AGENTS.md 已被超过 60,000 个开源项目和代理框架采用，包括 Amp、 Codex、 Cursor、 Devin、 Factory、 Gemini CLI、 Github Copilot、 Jules 和 VS Code 等——这反映出在代理进入生产环境时，对共享、供应商中立约定的广泛共识。

将 AGENTS.md 捐赠给 AAIF 可以确保：

• 开发者和企业有一种可移植的方式为代理提供所需的上下文；
• 该格式能在开放环境中演进，接受来自多种工具和社区的反馈；
• 随着采用增加，不会被单一公司主导方向。

除了 AGENTS.md 外， AAIF 的联合创始方还将各自贡献项目： Anthropic 的 Model Context Protocol ( MCP ) 和 Block 的 goose。

AAIF 能带来什么

AAIF 将成为一个中立的平台，在这里可以共同制定、治理并扩展代理互操作性标准。我们的目标很明确：建立一个共享且中立的平台，支持构建有利于开发者、企业和更广泛开源社区的代理系统。

AAIF 作为 Linux Foundation 之下的一个定向基金运作。 Linux Foundation 是一个具有数十年经验的非营利组织，曾管理 Linux Kernel、 Kubernetes、 Node.js 和 PyTorch 等关键开源项目，擅长提供中立治理、社区建设和长期可持续性保障，确保基础设施的发展不被单一公司所左右。

由 OpenAI、 Block、 Anthropic、 Google、 Microsoft、 Amazon、 Bloomberg 和 Cloudflare 联合发起的 Agentic AI Foundation，旨在通过共同投资、开放开发和社区主导的标准，让代理式 AI 的演进更加透明、协作且符合公共利益。

我们欢迎各类贡献者——工具构建者、研究人员、企业和个人开发者——参与并推动这项工作。了解更多关于 AGENTS.md 的信息并参与 AAIF，请访问 AAIF 的相关页面。

Today, OpenAI is co-founding the Agentic AI Foundation (AAIF)⁠ under the Linux Foundation, alongside Anthropic and Block, and with the support of Google, Microsoft, AWS, Bloomberg, and Cloudflare. The AAIF is designed to provide neutral stewardship for open, interoperable infrastructure as agentic AI systems move from experimentation into real-world production.

As part of this effort, we’re contributing AGENTS.md⁠—a simple, open format for providing agents with project-specific instructions and context—to the foundation to ensure long-term support and adoption across the community.

Developers are rapidly adopting AI to build more capable agentic systems—from coding assistants to workflow automation and customer service agents. In 2025, these systems have begun to move from prototypes into tools that handle real work in business and consumer settings.

We believe the transition from experimental agents to real-world systems will best work at scale if there are open standards that help make them interoperable. Open standards make agents safer, easier to build, and more portable across tools and platforms—and help prevent the ecosystem from fragmenting as this new category matures.

As more agents begin handling real responsibility, the cost of fragmentation increases. Without common conventions and neutral governance, agent development risks diverging into incompatible silos that limit portability, safety, and progress.

AAIF was created to address this moment.

Over the past year, OpenAI has made significant contributions to building agentic infrastructure in the open—working alongside developers, researchers, and other companies across the ecosystem to prove that open protocols can work in practice.

We’ve contributed core building blocks including the Agents SDK⁠, the Apps SDK⁠, and the Agentic Commerce Protocol⁠, as well as open-source efforts like our gpt-oss models⁠ and the Codex CLI⁠. The Codex CLI alone has now been used to help merge more than two million public pull requests on GitHub.

We’ve also been early adopters and core contributors to the Model Context Protocol (MCP), incorporating it as the foundation for connectors⁠ and apps in ChatGPT⁠. Just last week, we announced a collaboration with Anthropic and MCP-UI to extend the Apps SDK to all MCP developers through MCP Apps⁠—making it easier to build interactive, agent-powered experiences across platforms.

These efforts laid the groundwork for AAIF by demonstrating that open, interoperable infrastructure can gain real adoption across tools and communities.

AGENTS.md began as a practical solution to a familiar problem. Codex needed a predictable way to find project-specific instructions—such as coding conventions, build steps, and testing requirements—to help agents work safely and effectively in a codebase.

With AGENTS.md⁠, we introduced a lightweight Markdown file designed to live alongside README.md⁠. It offers a simple, interoperable standard that gives AI coding agents a consistent source of project-specific guidance needed to operate reliably across different repositories and toolchains. This markdown-based convention makes agent behavior far more predictable across diverse repositories and build systems.

What began as an internal solution quickly resonated across the ecosystem. Since its release in August 2025, AGENTS.md has been adopted by more than 60,000 open-source projects and agent frameworks including Amp, Codex, Cursor, Devin, Factory, Gemini CLI, Github Copilot, Jules and VS Code among others—reflecting growing alignment around shared, vendor-neutral conventions as agents enter production.

Contributing AGENTS.md to the AAIF ensures that:

• Developers and enterprises have a portable way to provide agents with the context they need
• The format can evolve in the open, with input from many tools and communities
• No single company controls its direction as adoption continues to grow

In addition to AGENTS.md⁠, the AAIF co-founders will each be contributing a project to the foundation: Anthropic’s Model Context Protocol (MCP)⁠ and Block’s goose⁠.

The AAIF provides a neutral home where agent interoperability standards can be developed, governed, and extended collaboratively. Our goal for the AAIF is simple: establish a shared, neutral platform for building agentic systems that benefit developers, enterprises, and the broader open-source community.

AAIF operates as a directed fund under the Linux Foundation, a non-profit organization with decades of experience stewarding critical open-source projects including the Linux Kernel, Kubernetes, Node.js, and PyTorch. The Linux Foundation brings proven expertise in neutral governance, community building, and maintaining long-term sustainability for widely adopted technologies—ensuring no single company controls the direction of foundational infrastructure.

Established by OpenAI, Block⁠, Anthropic⁠, Google⁠, Microsoft⁠, Amazon⁠ and Bloomberg⁠, and Cloudflare⁠, the Agentic AI Foundation is designed to help agentic AI evolve transparently, collaboratively, and in the public interest through shared investment, open development, and community-led standards.

We welcome contributors of all kinds—tool builders, researchers, enterprises, and individual developers—to help shape and extend this work. Learn more about AGENTS.md⁠ and get involved with the AAIF here⁠.

Generated by RSStT. The copyright belongs to the original author.

Source