HTML · ★ 14,315 · 🍴 858 · 📈 4,504 stars today
29 editorial diagram types for Claude Code. Self-contained HTML + SVG. No shadows, no Mermaid-slop.
中文介绍 为 Claude Code 提供 29 种编辑级图表类型，采用自包含 HTML+SVG，无阴影、不依赖 Mermaid 的粗糙默认样式。适合在智能体工作流中直接生成清晰、可定制的示意图。
Python · ★ 6,597 · 🍴 696 · 📈 727 stars today
Graph-Native Infrastructure for Context and Accountable AI Systems
中文介绍 面向上下文与可问责 AI 系统的图原生基础设施，用图结构组织和管理上下文，解决 AI 系统在复杂场景下的记忆、溯源和责任归属问题。
Python · ★ 168,994 · 🍴 20,129 · 📈 383 stars today
Public repository for Agent Skills
中文介绍 Anthropic 官方发布的 Agent Skills 公开仓库，用于定义和共享可复用的智能体技能，帮助开发者构建更强大的 Claude 自动化工作流。
Python · ★ 4,930 · 🍴 333 · 📈 768 stars today
14MB foundation model for tiny devices; phones, wearables, smart home, and robots.
中文介绍 仅 14MB 的端侧基础模型，专为手机、可穿戴设备、智能家居和机器人设计，让资源受限的小型设备也能本地运行 AI 推理。
Swift · ★ 9,836 · 🍴 663 · 📈 187 stars today
Fastest and only macOS Dictation app with on-device STT and custom trained AI enhancement model. A local Wispr Flow alternative. ⭐ helps a ton :) Windows & iOS waitlist open. Linux soon.
中文介绍 macOS 本地听写应用，采用端侧 STT 和自训练 AI 增强模型，号称最快且唯一的本地替代品，对标 Wispr Flow，主打隐私和低延迟。
Python · ★ 71,026 · 🍴 6,405 · 📈 354 stars today
Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.
中文介绍 本地 UI 工具，用于运行和训练 LLM 与扩散模型，支持 Qwen3.8、Kimi K3、Gemma 4、DeepSeek-V4、FLUX 等新模型，降低本地部署和微调门槛。
Rust · ★ 2,578 · 🍴 276 · 📈 1,180 stars today
Macro is a unified workspace for teams: email, chat, docs, tasks, agents, calls, and CRM — @-linked together with shared AI memory.
中文介绍 团队统一工作空间，整合邮件、聊天、文档、任务、Agent、通话和 CRM，所有内容通过 @ 链接串联，并共享 AI 记忆，提升协作效率。
Python · ★ 12,400 · 🍴 1,667 · 📈 166 stars today
holehe allows you to check if the mail is used on different sites like twitter, instagram and will retrieve information on sites with the forgotten password function.
中文介绍 检查邮箱是否在 Twitter、Instagram 等网站注册，并利用“忘记密码”功能获取关联信息，用于账号关联分析、隐私审计和社工测试。
Python · ★ 20,652 · 🍴 3,313 · 📈 278 stars today
SpiderFoot automates OSINT for threat intelligence and mapping your attack surface.
中文介绍 自动化 OSINT 工具，集成多种数据源，用于威胁情报收集和攻击面测绘，帮助安全团队快速发现暴露在外的资产和潜在风险。
Rust · ★ 1,193 · 🍴 107 · 📈 408 stars today
Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimization.
中文介绍 LLM 应用流量路由层，兼容 OpenAI 和 Anthropic 原生 API，可在不同模型和供应商之间灵活切换，方便做模型选型、基准测试和成本/性能优化。
TypeScript · ★ 6,562 · 🍴 598 · 📈 380 stars today
Open-source All in One AI agent workspace. Run any agent — Claude Code, Codex — across your tools (100+ integrations + MCP), apps, browser, and files, with shared memory. Built-in models or BYOK.
中文介绍 开源一体化 AI Agent 工作空间，可运行 Claude Code、Codex 等任意 Agent，支持 100+ 工具集成和 MCP，内置共享记忆，可用自带模型或 BYOK。
★ 45,694 · 🍴 3,295 · 📈 411 stars today
Agent skills for Obsidian. Teach your agent to use Obsidian CLI and open formats including Markdown, Bases, JSON Canvas.
中文介绍 为 Obsidian 设计的 Agent skills，教智能体使用 Obsidian CLI 及开放格式（Markdown、Bases、JSON Canvas），实现笔记和知识库的自动化操作。
Python · ★ 90,837 · 🍴 7,529 · 📈 204 stars today
Animation engine for explanatory math videos
中文介绍 数学讲解视频动画引擎，用 Python 编写，可精确控制几何图形和公式动画，广泛用于 3Blue1Brown 风格的科普视频和教育内容制作。
Shell · ★ 145,168 · 🍴 23,483 · 📈 762 stars today
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
中文介绍 一套完整的 AI 代理团队，包含前端专家、社区运营、创意注入等多种角色，每个 Agent 都有专属个性和流程，可用于自动化营销、社区管理和内容创作。
Python · ★ 8,902 · 🍴 1,406 · 📈 201 stars today
Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.
中文介绍 LTX-2 音频-视频生成模型的官方 Python 推理和 LoRA 训练包，支持本地生成音视频内容，并能通过 LoRA 微调定制风格。
TypeScript · ★ 5,382 · 🍴 574 · 📈 221 stars today
Desktop app to generate 3D models from images using local AI — runs entirely on your GPU
中文介绍 桌面应用，利用本地 AI 从图片生成 3D 模型，完全在 GPU 上运行，无需上传数据，保护隐私，适合设计师和开发者快速生成素材。
Go · ★ 87,997 · 🍴 10,350 · 📈 473 stars today
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
中文介绍 开源 RAG 引擎，融合检索增强生成与 Agent 能力，为 LLM 提供高质量上下文层，支持深度文档理解和智能问答，是大模型落地的重要工具。
👍 182
AI agents operate in persistent environments where early state changes can influence decisions far into the future. Unlike conventional language-model interactions, agent behavior is mediated through a shared state that is repeatedly modified and reused across long-horizon workflows. Current safety
中文介绍 提出 OpenART，通过开放式环境演化扩展智能体红队测试。针对智能体在持久环境中早期状态变化影响长期决策的安全风险，自动演化共享状态与任务场景，发现常规静态评测难以暴露的危险行为，提升安全评测覆盖度。
👍 181
After an older adult misses a medication dose, a software agent can send another reminder and an embodied agent can bring the medication. Yet neither explains whether the person forgot, is confused, has side effects, or deliberately refused, nor what support is appropriate. This reveals a structural
中文介绍 提出 ComBodied Agents 范式，将软件智能体与具身智能体结合，用于理解人类状态与意图。以老人漏服药物为例，现有系统只会提醒或送药，无法判断原因并提供恰当支持；该范式强调多模态感知、推理与主动帮助，实现人本智能。
👍 176
Turning a research idea into a complete paper requires more than text generation: the system must retrieve literature, design and execute experiments, revise claims according to evidence, produce publication-ready figures, and maintain consistency across a long generation process. We present Spark-t
中文介绍 提出 Spark-to-Paper，将研究想法生成完整论文的流程封装为可组合技能。系统自动完成文献检索、实验设计与执行、依据证据修订论断、生成出版级图表，并保持长流程各环节一致性，实现端到端学术论文生成。
👍 124
Agentic systems are increasingly expected to improve after deployment, yet single-entity self-evolution is often bounded by a static learning context, such as fixed tasks and feedback. This survey focuses on co-evolution in agentic systems, a multi-component form of self-evolution in which multiple
中文介绍 综述智能体系统中的共演化：多个组件在部署后相互适应、共同进化，突破单体自我演化受固定任务与反馈约束的局限。围绕多组件自演化的机制、类型与挑战展开梳理，为超越人类预设的自主演化系统提供研究框架。
👍 97
Recent work on distillation transfers the capabilities of large models to smaller ones often by updating the latter's parameters, through teacher forcing, on-policy distillation, and related training-time methods. In this paper, we ask whether such transfer can instead occur at test time. We study s
中文介绍 研究测试时的强到弱能力迁移：不更新小模型参数，而通过 harness 在推理阶段传递大模型能力，与训练时蒸馏互补。实验验证测试时蒸馏的可行性，为能力迁移提供免训练、低成本的新途径。
👍 73
AI models have achieved remarkable success across diverse domains, yet the mechanisms underlying their capabilities and the risks they may pose remain poorly understood. As AI development becomes faster and increasingly automated, mechanistic exploration remains largely manual, widening the gap betw
中文介绍 提出 Mechanist，将 AI 用作科学仪器，自动探索智能机制的内部原理。针对模型能力与风险机理仍不明确、机制研究高度手动且落后于 AI 开发速度的问题，自动化机制发现流程，缩小模型开发与理解之间的差距。
👍 72
Large Language Models (LLMs) increasingly act as agents whose procedural knowledge is stored in reusable skill packages and loaded at inference time. As skill libraries grow, a central challenge is to expose the smallest sufficient executable context under a limited context budget. Existing systems
中文介绍 提出 SkillZip，一种保持契约的图压缩方法，用于可扩展智能体技能库。在上下文预算有限时，通过压缩技能依赖图，暴露最小且足够的可执行上下文，解决技能库规模增长带来的推理加载成本问题。
👍 26
Self-improving coding agents that iteratively rewrite their own source code have demonstrated impressive performance on coding tasks. However, existing solutions generally derive self-modification from a single failure trajectory at a time, overlooking rich comparative signals available in the agent
中文介绍 提出 Mendel Gödel Machine，通过比较多条失败轨迹驱动编码智能体递归改写自身代码。不同于现有方法只从单条失败轨迹学习，该机制利用轨迹间的对比信号实现更稳健的自我改进，提升编码智能体的迭代优化效果。
👍 25
The rapid advancement of Large Language Models (LLMs) is revolutionizing AI for Games by enabling open-ended and fluid interactive storytelling. However, existing research has largely overlooked the critical challenge of maintaining long-horizon logical consistency and narrative integrity against un
中文介绍 提出基准测试，评估 LLM 智能体在开放式交互叙事中面对意外事件时，能否维持长期逻辑一致性与叙事完整性。现有研究较少关注该问题，本工作为 AI 玩游戏与互动故事生成提供一致性评测标准。
👍 17
Large language model (LLM) agents are increasingly deployed as personal assistants. Existing evaluations, however, mostly use short, self-contained requests in static environments. Everyday life assistance is different. A task runs for weeks rather than minutes. The world keeps changing while the ag
中文介绍 提出 VibeLifeBench，评测生活助手智能体在持续变化的真实世界中的主动性与持久性。现有基准多为静态环境下的短时请求，该基准要求任务跨越数周、环境不断变化，检验智能体长期规划与主动执行能力。
👍 13
Large language model evaluations typically focus on performance under nominal conditions, creating an illusion of capability where models comfortably walk a narrow, highly optimized generation corridor. In real-world deployments, however, complex system prompts, safety guardrails, and structural con
中文介绍 提出解码层禁忌测试，作为诊断 LLM 鲁棒性的压力测试。常规评测只看标准条件下的表现，该测试要求模型在复杂系统提示、安全护栏和结构约束下生成，暴露其脱离最佳生成通道后的脆弱点，补齐鲁棒性评估缺口。
👍 13
Self-evolving agents accumulate reusable skills by appending successful procedures and failure fixes. Over time, the same requirement is often restated in several branches, examples, and warnings, while common action sequences are copied rather than reused. The resulting skill becomes expensive to i
中文介绍 提出一种免评估的技能压缩方法 SkillZip：通过发现可复用结构，压缩自演化智能体不断累积的技能库。针对同一需求被重复表述、动作序列被复制等冗余，自动抽取公共结构，降低技能存储与推理开销。
👍 12
Long-horizon research agents solve open-ended tasks through iterative retrieval, aggregation, and synthesis, but context grows rapidly while the marginal value of additional evidence often declines. This leads to unnecessary token cost, higher latency, and noisier inputs for final report generation.
中文介绍 提出边际价值估计方法，用于提升深度研究智能体的效率。长程研究智能体在检索、聚合与综合中上下文快速膨胀，而新增证据价值递减；通过估计边际价值决定是否继续检索，减少 token 成本、延迟和最终报告噪声。
👍 12
We study reference-free post-training for multilingual machine translation with open large language models. Starting from the supervised-finetuned MiLMMT-46-v0.1 models, we apply Group Relative Policy Optimization (GRPO) with a reward that averages two reference-free quality estimation models and is
中文介绍 研究无参考后训练提升开源大模型的多语言机器翻译。在监督微调模型 MiLMMT-46-v0.1 上应用 GRPO，以两个无参考质量估计模型的平均得分作为奖励，不依赖参考译文即可优化翻译质量，扩展多语言能力。
👍 11
Long-document understanding often requires reasoning over many visually rich pages, making inference costly and prone to context rot. In this work, we propose InSight-doc, an agentic visual perception framework that treats visual resolution as an adaptive reasoning-time resource. InSight-doc starts
中文介绍 提出 InSight-doc，一种智能体视觉感知框架，将视觉分辨率视为可自适应调配的推理时资源。先低分辨率浏览、再按需放大关键区域，以处理多视觉页面的长文档理解，降低推理成本并缓解上下文腐烂问题。
👍 11
Visual document retrieval (VDR) is dominated by multi-billion-parameter models that are slow to index at full corpus scale and expensive to serve. Prior compression routes either train a smaller multi-vector encoder from scratch or distil only the query side; neither yields a compact single-vector r
中文介绍 提出 DistilVDR，通过双学生蒸馏得到紧凑端到端视觉文档检索器。现有主流 VDR 模型参数达数十亿，索引与推理成本高；此前压缩要么从头训练小模型、要么只蒸馏查询端，该方法产出低维单向量检索器，兼顾效率与质量。
👍 11
Large language models (LLMs) are increasingly deployed as mobile assistants, where a key challenge is leveraging personal information scattered across multiple applications (apps) to complete user instructions. However, due to the lack of dedicated benchmarks, their capabilities remain poorly unders
中文介绍 提出 SPIEval，评估 LLM 作为移动助手时，对散落在多个应用中的个人信息的利用能力。该基准填补专门评测空白，系统考查跨应用信息定位、整合与指令执行能力，为移动端个人助理研究提供标准测试集。
👍 10
World modeling is an unsettled field: architectures, training objectives, and state representations interact in complex ways, and no single recipe dominates across environments. This makes it an ideal testbed for AI coding agents acting as autonomous researchers--a setting in which the improvement d
中文介绍 提出 AutoWorldModel-Bench，以状态为中心的基准平台，用于自动化世界模型研究，并作为 AI 编码智能体担任自主研究员的测试床。世界模型架构、训练目标与状态表示组合复杂，该基准提供统一评测环境，促进自动改进。
👍 10
We present 360CityArena, a benchmark for evaluating the urban exploration capabilities of embodied agents within a photorealistic environment constructed from 360-degree videos. Existing outdoor benchmarks either lack sufficient photorealism or complexity, resulting in a considerable gap from real-w
中文介绍 提出 360CityArena，基于 360 度视频构建照片级真实城市环境，评测具身智能体的城市探索能力。现有室外基准在逼真度或复杂度上不足，离真实世界较远；该基准提供更接近实际场景的导航评测。
👍 9
Embodied agents are increasingly built as systems around foundation models, where performance depends not only on model weights but also on the skills, context, action interfaces, and execution harness surrounding the model. While supervised fine-tuning and reinforcement learning can adapt agents to
中文介绍 提出技能-执行框架演化方法，让具身智能体在部署后自我演化：同时优化技能、上下文、动作接口与执行框架，而不只调整模型权重。相比监督微调与强化学习，该方法从整体系统层面提升智能体在真实世界中的适应能力。
@thejessezhang · 85.7K 粉丝 · 270.4K 阅 · 510 赞 · 30 转
Two-thirds of our deployment work is now done autonomously by our own product (via Duet). This post is about why that is, our philosophy on building a product + deployment motion, and how that's
中文介绍 作者分享团队经验：三分之二部署工作已由自家产品 Duet 自主完成，并深入阐述其产品构建与部署理念，以及这套自动化体系的设计思路，讨论为何选择让产品自治部署而非人工介入。
@addyosmani · 408.3K 粉丝 · 175.5K 阅 · 509 赞 · 65 转
For much of human history, we've evaluated code quality via code review: someone reads what you wrote and makes sure it's clean, thoughtful, fast, understandable, and tests well. For agents, that
中文介绍 探讨智能体时代的代码质量审查：人类历史上靠人工代码审查，而对代理生成代码，传统标准不再适用，需要新的评估方式，转向“代理生产代码”的质量范式。
@nvidia · 2.6M 粉丝 · 59.5K 阅 · 548 赞 · 66 转
The new lightweight open model and routing library delivers greater control over AI, data and workflows across edge devices, PCs, workstations, data centers and the cloud. Source: NVIDIA Blog By Kari
中文介绍 NVIDIA 发布轻量开源模型 Nemotron 3.5 Lightning 与路由库 NeMo Switchyard，可提升 Agentic AI 的速度与智能，并支持从边缘设备、PC 到数据中心的灵活部署。
@GoogleCloudTech · 1.3M 粉丝 · 56.7K 阅 · 505 赞 · 76 转
Agent Plugins is an open, vendor-neutral standard for packaging Agent Skills and the MCP servers they depend on into one portable folder that any compatible client can load. Google is joining the
中文介绍 Google 加入 Agent Plugins 开放标准，将 Agent Skills 与其依赖的 MCP 服务器打包为一个可移植文件夹，任何兼容客户端均可直接加载，旨在构建无供应商锁定的 Agent 生态。
@dexhorthy · 30.0K 粉丝 · 42.9K 阅 · 635 赞 · 31 转
tl;dr make your agent converse visually instead of in walls of prose. Lighter and faster than HTML, good enough for most dev-work shaped problems. Coding agents are pretty much unreadable The
中文介绍 提出 /show-me 机制，让编码代理用紧凑视觉表示代替长篇文字，比 HTML 更轻量快速，足以应对多数开发问题，解决代理输出难以阅读的痛点，让代理“用图说话”。
@mattyp · 44.5K 粉丝 · 33.5K 阅 · 651 赞 · 48 转
I’m always chasing better tools. Notes apps, workflows, optimizations, and now, personal agents. It’s an easy trap to fall into - always building the “custom” thing and sacrificing the work as a
中文介绍 作者分享对个人代理工具的追逐与反思：从笔记应用到工作流，介绍 Grok Bot，提醒别掉进“总想构建定制工具”的陷阱，认为使用现成工具比自建更重要。
@ericzakariasson · 80.4K 粉丝 · 32.2K 阅 · 719 赞 · 52 转
Grok 4.6 is out! I've used it for a few weeks as my daily driver across the normal mix of coding and knowledge work, and built a few projects with it specifically to push on where it holds up. It's
中文介绍 Grok 4.6 发布，作者将其作为数周日常主力，覆盖编码与知识工作，并专门构建多个项目测试其能力边界，撰写实战指南，评估优势与不足。
@GoogleAIStudio · 191.7K 粉丝 · 30.7K 阅 · 566 赞 · 60 转
Today, we’re building on the progress of our widely used Flash series by introducing Gemini 3.7 Flash, our most intelligent workhorse model yet for coding and agents. This release comes just three
中文介绍 谷歌发布 Gemini 3.7 Flash，定位为面向编码与智能代理的最强“工作马”模型，在 Flash 系列广泛使用的基础上进一步智能化，主打高效实用，保持高频迭代节奏。
中文介绍 Hugging Face博客介绍Strands Agents、LeRobot与Hugging Face Storage Buckets的集成方案，实现从数据记录、模型训练到部署的一站式流程，帮助开发者更高效地构建机器人智能体。
中文介绍 DeepMind官方发布Gemini 3.7 Flash模型，作为Gemini系列的最新版本之一，目前官方仅公布发布消息，详细性能参数尚未公开。
The police-tech giant Flock is announcing today that it will change officers’ access to its nationwide network of license plate readers, in an apparent effort to quell a growing backlash and win back contracts lost amid concerns about mass surveillance and police abuse. Several changes aim directly
中文介绍 美国警用科技公司Flock宣布调整警官对其全国车牌读取器网络的访问权限，以回应日益强烈的大规模监控与警务滥权担忧，并希望赢回因此失去的合同。
Learn how startups use GPT-5.6 to build faster, more cost-efficient AI agents with smarter model selection and new Responses API capabilities.
中文介绍 OpenAI发布《GPT-5.6开发者指南》，介绍初创公司如何利用该模型构建更快、更具成本效益的AI智能体，包括更智能的模型选择以及新的Responses API能力。
Preview Ultrafast, a new OpenAI API service tier that runs GPT-5.6 Sol up to 14× faster. Powered by Cerebras, it delivers up to 750 output tokens per second.
中文介绍 OpenAI预览新的API服务层级Ultrafast，由Cerebras提供算力，运行GPT-5.6 Sol速度提升最高14倍，输出速率达每秒750个token。
OpenAI appoints Dali Rajic as Chief Revenue Officer to lead its global revenue organization and help businesses realize the full value of AI.
中文介绍 OpenAI任命Dali Rajic为首席营收官，负责领导全球营收组织，帮助各类企业充分实现AI的全部价值。
When we set out to talk to kids about artificial intelligence, we thought we knew what we’d hear. We expected some to tell us they were using it to cheat a little, the way Millennials and Gen Xers opened up CliffsNotes or programmed formulas into their TI-82s, and others to share inspiring ways they
中文介绍 MIT Tech Review通过访谈了解孩子们对人工智能的真实看法，原本预期一些孩子会承认用AI作弊或分享使用经历，但受访者的回答与预期不尽相同。
AI teammate category just had its most significant new entrant yet
中文介绍 Latent Space报道，SpaceXAI发布Grok 4.6与Grok @Bot，称其为AI队友类别迄今最重要的新进入者。
中文介绍 Hugging Face团队分享复现ICML会议2200篇论文的经验，探讨学术复现中的挑战与发现，为AI研究可重复性提供参考。
中文介绍 TLDR AI汇总当日重要动态：Claude推出Chrome Cowork协作功能，Grok 4.6发布，DeepSeek推出v4-Pro-0813模型。
Business and technology leaders need no convincing that the time of agentic AI is here. Organizations are rapidly adopting agents, and few executives doubt the technology’s potential to transform work. But many organizations find that realizing the desired return on investment (ROI) from AI hinges o
中文介绍 文章探讨如何用可信数据扩展AI智能体。企业正快速采用智能体，但许多组织发现，要实现预期投资回报，仍需解决数据可信度等挑战。
中文介绍 Hugging Face与艾伦人工智能研究所（AI2）推出OlmoEarth嵌入功能，支持从OlmoEarth Studio导出自定义嵌入向量，用于下游分析任务。
Introducing sign-language-to-text (SL2T), our breakthrough model powering new sign language features for Deaf and hard of hearing users.
中文介绍 DeepMind推出突破性手语转文本模型SL2T，将驱动面向聋人和听障用户的新型手语功能，帮助他们更便捷地进行交流。
中文介绍 Liquid AI发布LFM2.5-VL-3B模型，为边缘设备提供更强、更快的视觉能力，优化低资源环境下的多模态推理性能。
Speculative Decoding by any other name would distil as sweet
中文介绍 Latent Space探讨如何窃取推理轨迹，文章指出推测解码等方法本质上与知识蒸馏异曲同工，可被用于提取推理过程。
美股整体技术面偏强：SPY 距 52 周高点仅 -0.19%、QQQ 近 5 日涨 2.44%，二者均站上主要均线，RSI 位于 60-67 区间，MACD 红柱延续；但 MSFT 已现 RSI 超买（71.8），TSLA 虽出现 MACD 金叉却仍处空头排列。加密市场情绪偏冷：恐慌贪婪指数 29（恐慌），总市值 2.26T USD，BTC 主导率 56.2%、ETH 10.1%；BTC 低于 200 日均线且 MACD 死叉，ETH 围绕 50 日均线震荡，整体中性偏弱。中概股短线承压明显，PDD、京东、腾讯均出现 MACD 死叉，RSI 回落至 39-42 区间，5 日跌幅介于 7%-10.7%。商品外汇方面，黄金 5 日涨 4.08%、RSI 68.6 偏强，原油 MACD 金叉但日内回落，美元兑人民币 RSI 29.3 超卖并贴近 52 周低点。
价格 777.88 距 52 周高点仅 -0.19%，近 5 日涨 1.21%，站上 SMA20（754.54）、SMA50（748.48）和 SMA200（705），呈多头排列；RSI14 为 67.4 未超买，MACD（8.26）位于信号线（5.69）上方，动量保持正向。整体技术状态偏上行。
原油现价 81.15，5 日累计涨 4.99%，站上 SMA50（79.82）与 SMA200（76.73），触发 MACD 金叉（MACD 0.22 高于信号线 0.17），且信号标注为多头排列；不过日内 -2.55%，且价格仍略低于 SMA20（82.50），短线存在回踩确认需求。整体技术面偏上行。
美元兑人民币报 6.74，处于 52 周低位附近，RSI14 仅 29.3 进入超卖区，MACD（-0.0092）低于信号线（-0.0079），价格低于 SMA20/50/200，空头排列明确；近 1 日与 5 日分别下跌 0.13% 与 0.20%，下行趋势未见反转信号。技术状态偏下行。
京东现价 29.30，近 1 日跌 7.31%、5 日跌 10.70%，RSI14 降至 39.0，MACD（0.71）低于信号线（1.03）并触发死叉；价格跌破 SMA20（31.60），贴近 SMA50（29.21）与 SMA200（29.41），短线支撑薄弱。整体技术面偏下行。
阿里巴巴现价 122.16，处于 SMA20（121.37）附近，高于 SMA50（113.83）但低于 SMA200（139.23），RSI14 为 52.7 属中性区；MACD（3.85）与信号线（3.72）贴近，红柱微弱，近 1 日跌 2.44%、5 日跌 3.67%，上下动能均不突出。技术状态中性。
VIX 恐慌指数
10Y 美债收益率 (%)
美元指数 DXY
S&P 500 ETF
Nasdaq 100 ETF
Apple
Microsoft
Nvidia
Alphabet
Tesla
Meta
Bitcoin
Ethereum
Solana
阿里巴巴 (BABA)
拼多多 (PDD)
京东 (JD)
腾讯控股 (0700.HK)
黄金期货
WTI 原油期货
美元 / 人民币
本报告仅基于公开行情数据的技术指标读数，不构成任何投资建议。技术指标存在滞后性，过去走势不代表未来表现，市场可能随时变化。以上内容仅供技术指标解读参考。
Latest Israeli attacks kill police official Jamal Abu Kamil in Gaza City and another person in Khan Younis.
Thousands of sailors on the USS Abraham Lincoln have reportedly faced food shortages and broken plumbing, with some considering jumping overboard.
PM says he also spoke with the US president about Aukus, which ‘remains full steam ahead’. Follow today’s news live Get our breaking news email, free app or daily news podcast Albanese says gambling inducements are ‘over the top’ Albanese said gambling inducements are “over the top and need to be wo
Prime Minister Sanae Takaichi of Japan has been a public critic of Mr. Putin, who has denounced her country’s sanctions against Russia over the Ukraine war.
A preliminary report by US investigators says an engine fan blade broke on last month's flight from Greece to Germany.
Emergency services said they were working at the scene of a derailment near Lewes, in East Sussex, on Thursday. Eighteen other people were less badly hurt.
Decrees restricting women’s rights to study, work, travel and act independently now threaten to damage Afghanistan permanently, experts say.
Supporters of Andrew and Tristan Tate gathered outside a Miami detention centre ahead of their bail hearing.
U.S. ambassador to Israel Mike Huckabee called settlers besieging a West Bank home of an American-Palestinian family "terrorists."
Israeli drone strikes killed at least 2 people and wounded several others in Gaza on Thursday.
Colombia denies politics played a role after Mexico said its military rescuers were refused entry.
The economy dominates voter concerns in the southern African country.
Trustees condemn Kennedy Center’s move to re-add Trump’s name and shut down for extensive renovations.
At least six homes were set alight by an extensive wildfire in Stourbridge
A powerful explosion tore through a munitions faction south of the Italian capital Rome on Thursday.
After the PM announced a crackdown on subscription traps, readers share how they got into and out of unwanted plans.
中文摘要 英国首相宣布打击“订阅陷阱”后，读者分享如何陷入并退出不想要的订阅计划。一名读者成功获得89英镑退款，文章还介绍了取消订阅的实用方法，帮助消费者避免被自动续费困扰。
Rising prices are fueling a backlash against the government even as they signal a new era of economic dynamism after years of deflation. Bloomberg's Shery Ahn has more. (Source: Bloomberg)
中文摘要 日本通胀悖论正催生赢家和输家。物价上涨虽引发民众对政府的不满，却标志着通缩多年后经济活力的新时代。彭博社记者Shery Ahn报道称，通胀带来的影响呈现分化态势。
Three carriages on the Southern Rail service are flipped on to their side
中文摘要 英国一列南方铁路公司的列车发生脱轨事故，三节车厢翻倒侧翻，造成20人受伤。事故原因尚在调查中，相关部门已介入处理。
Asian stocks were poised to extend gains Friday as further evidence of moderating US inflation and a pullback in oil prices reinforced bets that the Federal Reserve will refrain from raising interest rates next month.
中文摘要 亚洲股市周五有望延续涨势，因美国通胀进一步放缓、油价回落，强化了市场对美联储下月将维持利率不变的押注。投资者情绪受到提振。
Reddit Inc. will join the S&P 500 next week as part of an off-cycle change, S&P Dow Jones Indices said Thursday, sending the social networking platform’s shares surging more than 10% after the bell.
中文摘要 标普道琼斯指数公司宣布，Reddit将于本月晚些时候以非例行调整方式被纳入标普500指数。消息公布后，该社交媒体平台盘后股价大涨逾10%。
Oil held a decline as traders monitored efforts toward a deal that could reopen the Strait of Hormuz, while fresh attacks on tankers and energy infrastructure kept the market on edge.
中文摘要 原油价格延续跌势，交易员密切关注可能重启霍尔木兹海峡通行的协议进展，同时针对油轮和能源基础设施的新一轮袭击令市场保持警惕。
A new book 'What Happened to Liberal Democracy?' by Nobel Laureate Daron Acemoglu examines the evolution of liberalism, and its future amid rapidly advancing technologies like AI. He joins Buisnessweek Daily to discuss why he wrote this book and how social media and online connections influence what
中文摘要 诺贝尔经济学奖得主达龙·阿杰姆奥卢在新书《自由民主怎么了？》中探讨自由主义演变及其在人工智能等快速技术变革下的未来。他认为，社交媒体和在线联系正将自由民主推向“失去”的边缘。
Some traders are reeling from heavy losses after a brutal correction in South Korea's stock market.
中文摘要 韩国股市经历剧烈波动，部分投资者损失惨重。一名投资者表示一个月内亏损1.4万美元。市场出现残酷回调，交易者正承受巨额亏损。
The Securities and Exchange Commission canceled a Friday meeting where the agency was expected to unveil new plans for digital assets as landmark crypto legislation remains stalled in Congress.
中文摘要 美国证券交易委员会取消了原定周五举行的会议，该会议本拟公布数字资产监管新计划，这是加密货币行业遭遇的最新挫折。同时，国会层面的加密立法仍处于停滞状态。
Kalshi was told by a judge to stop offering most of its prediction market contracts in Washington state, after regulators said they likely constituted an illegal gambling operation.
中文摘要 一名法官裁定，Kalshi必须停止在华盛顿州提供大部分预测市场合约。监管机构认为，这些合约可能构成非法赌博操作。此举对Kalshi的预测市场业务造成打击。
Also in today’s newsletter: Anthropic investors bet on $2tn valuation and Sanae Takaichi slams Vladimir Putin’s visit to disputed Pacific islands
中文摘要 今日FT简报：印度最大企业集团控制权争夺战成为焦点；Anthropic投资者押注其估值达2万亿美元；日本高市早苗批评普京访问有争议的太平洋岛屿。
It’s tough to measure the economy’s true potential
中文摘要 本文探讨衡量经济真实潜力的难度，指出传统经济指标存在种种局限，难以准确反映经济的潜在增长动力。
7 回复 · 程序员 节点
6 回复 · 程序员 节点
49 回复 · Apple 节点
97 回复 · Linux 节点
20 回复 · Linux 节点
12 回复 · Apple 节点
17 回复 · Apple 节点
39 回复 · Apple 节点
14 回复 · Apple 节点
6 回复 · Apple 节点
该源今日无内容。