Tag
Gaowei Chang（ANP开源社区创始人）深入解读Meta新发布的PAP（Personal Agent Protocol）规范，通过酒店预订场景说明其服务发现、用户授权与作用域控制机制，并对比MCP、A2A和ANP的设计差异。
作者分析近期行业动态，指出AI前沿模型频繁更替、共享格式（如SKILL.md、AGENTS.md）正在让配置变得可迁移，因此用户没有理由忠于单一助手，而真正有价值的将是保存跨模型工作记录的中间层。
作者观察到 MCP 生态的扩张热潮——各家 SaaS 公司纷纷推出自己的 agent 集成接口——并质疑这种每应用一个集成的模式能否规模化，呼吁行业转向共享的统一基础设施层，而非让每个企业维护几十个独立的 agent 门面。
Google的A2A协议从Linux基金会转入专注于Agentic AI的AAIF基金会，与MCP同属一个中立治理机构，成员已超250家，涵盖Google、Microsoft、Anthropic、OpenAI等，有望推动跨厂商智能体互操作标准落地。
分析 Meta 与 Sierra 联合推出的 Personal Agent Protocol (PAP)，探讨其解决个人代理代表用户与企业交互的身份与授权问题，并质疑是否真的需要新协议而非复用 MCP。
Bret Taylor announced Personal Agent Protocol, an open standard developed by Meta and Sierra along with partners including Shopify, Stripe, Walmart, and Genesys, defining how personal AI agents interact with businesses. Tim O'Reilly praised the move as a step toward user empowerment over platform lock-in.
FamRelay 是一个在不同 AI 助手之间共享事实的中继层，用户可审批要发送的信息，实现家庭成员或多账号助手（如 Claude 与 ChatGPT）之间的受控信息同步。
A blog essay arguing that plain text remains one of computing's most durable and trustworthy technologies, thanks to its simplicity, portability, human readability, and seamless integration with decades-old Unix tooling like grep, sed, and awk.
Qiniu Cloud's Lao Xu team open-sourced LLGo, an LLVM-based Go compiler that eliminates cgo overhead by enabling zero-cost native C ABI calls and integrating C, Python, and C++ ecosystems into Go, with support for disabling GC and targets ranging from games to Llama 2 inference and WebAssembly.
Kody is a software platform that enables AI agents to share resources like memory, secrets, and packages, improving interoperability and reducing the need to retrain agents in each session.
The article presents an architectural mediation approach using the Model Context Protocol to enable controlled interactions between LLM agents and data spaces, ensuring compliance with governance and interoperability standards.
ANP 1.2 is officially released, focusing on enabling different DID systems to join the same agent network, maintaining identity continuity after key rotation, and supporting secure cross-device communication.
The author proposes a local-first, distributed model for AI agents where users own persistent agents that can operate locally, negotiate with other agents, and interact across environments with structured actions and an agent economy.
A discussion about enabling AI assistants to communicate across different platforms, referencing Tincan, an open group chat tool for coordination.
German defense company Rheinmetall has open-sourced its Battlesuite connected weapon system protocol, releasing the onboardapi library and middleware for seamless communication between sensor systems and software components, built on the DDS standard.
PythonCall.jl is a library that enables seamless interoperability between Python and Julia, allowing code and data to be exchanged efficiently with symmetric interfaces and fast non-copying conversions.
This paper proposes a machine learning API specification for openEO to integrate ML workflows into Earth Observation data cubes, enhancing reproducibility and cross-platform interoperability. It demonstrates feasibility with prototype implementations in R and Python and use cases in crop type mapping and deep learning.
Code discoveries in iOS 27 and macOS Golden Gate reveal Apple has designed Siri to integrate with third-party AI models like Claude and ChatGPT, potentially allowing its AI to be swapped out.
This article explores the limitations of personal AI agents in interoperability, noting their inability to coordinate with other agents or systems, and questions whether current technologies can reliably address this gap.
This paper investigates portable semantics and negative transfer in latent communication between language-model cells, showing that independently trained societies share semantic interfaces but not raw language, and that inherited interfaces can cause severe negative transfer.