每日简报

2026-06-27

← 历史归档

simplex-chat/simplex-chat

Haskell · ★ 12,492 · 🍴 707 · 📈 191 stars today

SimpleX - the first messaging network operating without user identifiers of any kind - 100% private by design! iOS, Android and desktop apps 📱!

中文介绍 SimpleX 是首个无需用户标识的端到端加密消息网络,100% 隐私设计。基于去中心化架构,不依赖手机号或用户名,支持 iOS、Android 和桌面端。适合对隐私要求极高的个人或团队通信场景。

google-labs-code/design.md

TypeScript · ★ 21,179 · 🍴 1,717 · 📈 2,319 stars today

A format specification for describing a visual identity to coding agents. DESIGN.md gives agents a persistent, structured understanding of a design system.

中文介绍 DESIGN.md 是描述视觉设计系统的格式规范,为编码代理提供持久、结构化的设计理解。开发者可定义颜色、布局、组件等,使 AI 代理能遵循统一设计语言生成代码,适用于前端开发和设计系统维护。

commaai/openpilot

Python · ★ 61,770 · 🍴 11,047 · 📈 67 stars today

openpilot is an operating system for robotics. Currently, it upgrades the driver assistance system on 300+ supported cars.

中文介绍 openpilot 是机器人操作系统,当前升级 300 多款车型的辅助驾驶系统。采用计算机视觉和深度学习实现自适应巡航、车道保持等功能。适合汽车修改爱好者和自动驾驶研究者,在实际道路上测试和改进驾驶辅助能力。

kunchenguid/no-mistakes

Go · ★ 3,402 · 🍴 205 · 📈 412 stars today

git push no-mistakes

中文介绍 确保 git push 无差错的辅助工具。通过自动化检查(如代码格式、测试、lint)防止错误提交,简化开发者流程。适合希望减少推送失误的 Git 用户。

grafana/grafana

TypeScript · ★ 74,881 · 🍴 14,125 · 📈 17 stars today

The open and composable observability and data visualization platform. Visualize metrics, logs, and traces from multiple sources like Prometheus, Loki, Elasticsearch, InfluxDB, Postgres and many more.

中文介绍 Grafana 是开源可观测性与数据可视化平台,支持 Prometheus、Loki、Elasticsearch、InfluxDB 等多种数据源。提供仪表盘、告警和查询功能,用于监控指标、日志和链路追踪。适合运维、开发人员统一监控 IT 基础设施和应用性能。

ripienaar/free-for-dev

HTML · ★ 123,713 · 🍴 13,050 · 📈 137 stars today

A list of SaaS, PaaS and IaaS offerings that have free tiers of interest to devops and infradev

中文介绍 DevOps 和基础设施开发者可用的免费层 SaaS/PaaS/IaaS 服务清单,涵盖托管、数据库、监控、CI/CD 等类别。持续更新,帮助个人和小团队低成本搭建开发环境或学习新技术。

opendatalab/MinerU

Python · ★ 70,384 · 🍴 5,937 · 📈 944 stars today

Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

中文介绍 MinerU 将 PDF、Office 等复杂文档转换为 LLM 可直接使用的 Markdown 或 JSON 格式,支持嵌入表格、图片、公式等元素。适用于构建知识库、RAG 管道或 Agent 工作流,减少文档预处理工作量。

alchaincyf/zhangxuefeng-skill

★ 9,242 · 🍴 2,560 · 📈 185 stars today

张雪峰.skill — 张雪峰的认知操作系统。高考志愿/考研/职业规划的实战思维框架。由女娲.skill生成。

中文介绍 张雪峰认知操作系统,基于女娲.skill 构建的实战思维框架。涵盖高考志愿、考研、职业规划等决策场景,提供结构化的分析方法和策略。适合学生、求职者及教育咨询从业者使用。

mauriceboe/TREK

TypeScript · ★ 7,629 · 🍴 647 · 📈 1,063 stars today

A self-hosted travel/trip planner with real-time collaboration, interactive maps, PWA support, SSO, budgets, packing lists, and more.

中文介绍 TREK 是自托管的旅行计划工具,支持实时协作、交互式地图、PWA、单点登录、预算管理和打包清单。适合多人共同规划行程,数据自托管保障隐私。可用于个人或小团队旅行组织。

xbtlin/ai-berkshire

Python · ★ 3,086 · 🍴 454 · 📈 1,270 stars today

AI 时代的伯克希尔:基于 Claude Code 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built on Claude Code. 4 masters' methodologies + multi-agent adversarial analysis.

中文介绍 基于 Claude Code 的价值投资研究框架,集成巴菲特、芒格、段永平、李录四位大师方法论。通过多 Agent 对抗研究分析股票,生成投资报告。适合价值投资者利用 AI 辅助决策。

calesthio/OpenMontage

Python · ★ 23,559 · 🍴 2,615 · 📈 1,674 stars today

World's first open-source, agentic video production system. 12 pipelines, 52 tools, 500+ agent skills. Turn your AI coding assistant into a full video production studio.

中文介绍 OpenMontage 是首个开源 Agentic 视频制作系统,包含 12 个管道、52 个工具和 500 多项 Agent 技能。可将 AI 编码助手转变为完整视频制作工作室,支持脚本生成、素材剪辑、特效合成等流程。适合内容创作者和开发者自动化视频生产。

aws/agent-toolkit-for-aws

Python · ★ 1,344 · 🍴 120 · 📈 238 stars today

Official, AWS-supported MCP servers, skills, and plugins to help AI agents build on AWS

中文介绍 AWS 官方支持的 MCP 服务器、技能和插件集合,帮助 AI 代理在 AWS 上构建应用。提供对 AWS 服务的标准化接口,使代理能管理资源、部署服务、调试错误。适合在 AWS 环境中使用 AI 自动化运维和开发的团队。

NanmiCoder/MediaCrawler

Python · ★ 53,340 · 🍴 10,972 · 📈 640 stars today

小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 | 评论爬虫、微博帖子 | 评论爬虫、百度贴吧帖子 | 百度贴吧评论回复爬虫 | 知乎问答文章|评论爬虫

中文介绍 多平台社交媒体爬虫,支持小红书、抖音、快手、B 站、微博、百度贴吧、知乎等内容的帖子与评论抓取。提供统一接口,可用于数据采集、舆情分析、内容研究等场景。适合数据科学家和市场研究人员。

garrytan/gstack

TypeScript · ★ 116,604 · 🍴 17,303 · 📈 919 stars today

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

中文介绍 Garry Tan 的 Claude Code 设置,集成 23 个工具模拟 CEO、设计师、工程经理、发布经理、文档工程师和 QA 角色。开发者可直接使用这套工作流,提升代码生成和项目管理效率。适合通过 AI 加速软件开发流程的团队。

IceWhaleTech/CasaOS

Go · ★ 35,343 · 🍴 2,024 · 📈 612 stars today

CasaOS - A simple, easy-to-use, elegant open-source Personal Cloud system.

中文介绍 CasaOS 是简单易用的开源个人云系统,提供文件管理、应用商店、远程访问等功能。基于容器技术,适合家庭用户或小型团队在 NAS 或旧 PC 上搭建私有云盘和轻量服务器。

JCodesMore/ai-website-cloner-template

TypeScript · ★ 21,329 · 🍴 3,094 · 📈 1,076 stars today

Clone any website with one command using AI coding agents

中文介绍 用 AI 编码代理一键克隆任意网站的工具。输入目标 URL,代理自动下载前端资源并生成可运行副本,适用于快速搭建原型或学习前端架构。适合开发者和设计师。

Panniantong/Agent-Reach

Python · ★ 42,299 · 🍴 3,356 · 📈 1,164 stars today

Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

中文介绍 Agent-Reach 让 AI 代理通过 CLI 读取和搜索 Twitter、Reddit、YouTube、GitHub、Bilibili、小红书等网站,无需 API 费用。给代理“浏览器”能力,支持构建互联网感知的自动化工具,适合需要全网数据获取的 AI 应用。

Are We Ready For An Agent-Native Memory System?

👍 100

Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, update, consolidation, and dynamic lifecycle governance throughout agent execution. Despite this evolutio

中文介绍 研究LLM智能体的记忆系统,从简单检索增强演进为支持持久存储、更新、整合和生命周期管理的数据管理系统。提出原生智能体记忆架构,实现高效的信息治理。

ShutterMuse: Capture-Time Photography Guidance with MLLMs

👍 41

Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject-side recommendations, leaving the capture-time guidance capabilities of multimodal large language m

中文介绍 针对真实摄影缺乏拍摄时引导的问题,提出ShutterMuse,利用多模态大语言模型(MLLM)在拍摄时刻提供相机取景和主体姿态建议。现有裁剪基准仅关注事后预测,该方法填补了拍摄时引导的空白。

ViQ: Text-Aligned Visual Quantized Representations at Any Resolution

👍 34

A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-lev

中文介绍 提出ViQ,一种文本对齐的视觉量化表示方法,支持任意分辨率下的离散视觉表示。在保持低层细节的同时实现统一的多模态建模,缓解信息损失问题。

OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning

👍 33

Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet

中文介绍 针对基于结果的强化学习在语言智能体训练中仅有稀疏轨迹奖励的问题,提出OPID(在策略技能蒸馏)。通过密集的token级自蒸馏,结合强化学习稳定优化,提升中间决策的监督信号。

Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation

👍 32

While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation

中文介绍 指出文本到图像模型在处理真实世界请求时存在「上下文鸿沟」:用户意图隐式或不完整。提出Qwen-Image-Agent,通过上下文理解和动态知识增强,弥合用户输入与生成条件之间的差距。

MVTrack4Gen: Multi-View Point Tracking as Geometric Supervision for 4D Video Generation

👍 32

Synthesizing a novel-view video from a monocular reference video along a target camera trajectory requires both geometric consistency and motion fidelity with respect to the reference video. Existing methods based on explicit 3D representations are limited by the accuracy of off-the-shelf reconstruc

中文介绍 提出MVTrack4Gen,利用多视角点跟踪作为几何监督,从单目参考视频生成新视角视频。无需显式3D重建,通过跨视角点轨迹约束保持几何一致性和运动保真度。

The Verification Horizon: No Silver Bullet for Coding Agent Rewards

👍 30

A classical intuition holds that verifying a solution is easier than producing one. For today's coding agents, this intuition is being inverted: as foundation models develop stronger reasoning capabilities and engineering harnesses grow more sophisticated, generating complex candidate solutions is n

中文介绍 挑战「验证比生成更容易」的传统直觉,指出在编码智能体中,随着模型推理能力增强,生成复杂候选解变得容易,而确实验证其正确性反而成为瓶颈。提出「验证地平线」概念。

UnityShots: Memory-Driven Multi-Shot Audio-Video Generation with Boundary-Aware Gating

👍 23

Generating a coherent multi-shot video requires structured cross-shot memory. Subject appearance, scene context, and speaker identity must persist across cuts. Existing approaches either train end-to-end over fixed-length sequences and cannot scale, generate shot-by-shot with memory banks that grow

中文介绍 提出UnityShots,用于音频-视频联合多镜头生成。通过记忆驱动和边界感知门控机制,保持跨镜头的主体外观、场景上下文和说话者身份一致性,解决现有方法无法处理长序列或记忆爆炸的问题。

V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoning

👍 23

Fine-grained visual reasoning requires multimodal large language models (MLLMs) to identify task-relevant visual evidence and ground their reasoning in local image regions. Existing agentic methods typically rely on reinforcement learning with verifiable rewards or supervised fine-tuning on large-sc

中文介绍 提出V-Zero,用于细粒度视觉推理,无需答案标签即可进行在策略蒸馏。通过对比证据门控机制,让多模态大语言模型(MLLM)自主定位相关视觉区域并推理,避免依赖大规模人工标注。

JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting

👍 22

Speculative decoding (SD) accelerates autoregressive Large Language Models (LLMs) by drafting multiple tokens and verifying them in parallel, but it faces a scaling limitation: increasing the draft budget improves speed only when acceptance remains high and drafting overhead stays low. This ceiling

中文介绍 提出JetSpec,打破推测解码中扩展预算导致收益递减的瓶颈。通过并行树草稿机制,一次性生成多个候选token并并行验证,在保持高接受率的同时降低草稿开销,显著提升加速比。

GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents

👍 21

Computer-use agents can execute software tasks through either graphical interfaces or programmatic command interfaces, but existing evaluations confound interaction modality with differences in tasks, initial states, verifiers, and permitted actions. We introduce a matched execution-layer benchmark

中文介绍 构建匹配的执行层基准,系统比较GUI和CLI两种交互模式下计算机使用智能体的性能瓶颈。消除任务、初始状态和验证器的差异,揭示不同模态下的执行效率差异。

The Hitchhiker's Guide to Agentic AI: From Foundations to Systems

👍 13

The Hitchhiker's Guide to Agentic AI is a comprehensive practitioner's reference for building autonomous AI systems. The book covers the full stack from first principles to production deployment, organized around a central thesis: building great agentic systems requires understanding every layer of

中文介绍 一份全面的实践指南,覆盖构建自主AI系统的全栈知识,从基本原理到生产部署。核心思想:构建优秀智能体系统需要理解每个技术层的作用。

Fast LeWorldModel

👍 11

Joint-Embedding Predictive Architectures (JEPAs), including recent LeWorldModel (LeWM), have become a promising foundation for reconstruction-free visual world models. For visual planning, however, LeWM evaluates candidate action sequences by repeatedly applying a local one-step latent transition mo

中文介绍 提出Fast LeWorldModel,加速基于联合嵌入预测架构(JEPA)的视觉世界模型。通过改进候选动作序列的评估方式,避免重复进行单步潜变量预测,提高视觉规划速度。

Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It

👍 10

Tool use enables large language models (LLMs) to perform complex tasks, and recent agentic reinforcement learning (RL) methods show promise for enhancing model capabilities. However, RL alone often leads to instability or limited gains in tool-use tasks. In our experiments, some models exhibit catas

中文介绍 分析工具使用强化学习(RL)中模型会灾难性崩溃的原因,发现纯粹RL导致不稳定。提出引入监督信号(如专家演示)来稳定训练,避免能力崩塌并提升工具使用效果。

Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments

👍 9

As agentic systems continue to evolve and are widely deployed in real-world scenarios, there is a growing demand to faithfully evaluate their capabilities. However, current benchmarks are typically built on popular applications with relatively simple tasks and focus on a narrow set of capabilities w

中文介绍 重新评估智能体在超越熟悉环境时的能力。现有基准多基于简单任务和流行应用,该工作设计更严苛的测试场景,揭示当前智能体在复杂真实场景下的能力局限。

Autodata: An agentic data scientist to create high quality synthetic data

👍 9

We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a s

中文介绍 提出Autodata,一种让AI智能体扮演数据科学家、自动构建高质量合成训练和评估数据的方法。通过元优化训练数据科学家智能体,使其学会生成更优数据,用于下游模型训练。

LISA: Likelihood Score Alignment for Visual-condition Controllable Generation

👍 8

The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the sid

中文介绍 提出LISA,针对视觉条件可控生成中双分支范式(侧网络编码条件+融合到主网络)的旁路特征作用进行分析。通过似然得分对齐,更有效地利用旁路特征提升生成质量。

Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do

👍 8

Chain-of-Thought (CoT) has become a standard method for improving reasoning capabilities in large language models (LLMs) by eliciting step-by-step thinking, but its effectiveness in multimodal tasks remains unclear. In this paper, we aim to systematically investigate the key question: What can multi

中文介绍 系统研究多模态思维链(CoT)推理能力边界。发现CoT在视觉任务中并不总是有效,尤其在需要精确空间关系或细节推理时,反而可能引入幻觉。揭示其能做什么和不能做什么。

In-Context World Modeling for Robotic Control

👍 6

Modern Vision-Language-Action (VLA) models often fail to generalize to novel setups, such as altered camera viewpoints or robot morphologies, because they are typically conditioned only on current observations and language instructions. By ignoring the underlying system configuration as a variable,

中文介绍 提出上下文世界模型,让视觉-语言-动作(VLA)模型在控制中隐式建模系统配置(如相机视角、机器人形态)。通过上下文推理适应新配置,提升零样本泛化能力。

Confidence-Aware Tool Orchestration for Robust Video Understanding

👍 6

Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world em

中文介绍 指出视频推理模型存在「盲目信任问题」:对运动模糊、眩光等扰动下的帧同样信任,导致准确率下降15-30%。提出置信度感知的工具编排方法,动态选择可靠帧进行推理。

Introducing computer use in Gemini 3.5 Flash

@GoogleAIStudio · 179.4K 粉丝 · 41.0K 阅 · 605 赞 · 57 转

Computer use is now a built-in tool supported in Gemini 3.5 Flash, delivering our best performance yet for agentic computer use tasks. Previously only available as a standalone Gemini 2.5 computer use

中文介绍 Google AI Studio 宣布 Gemini 3.5 Flash 内置 computer use 工具,支持浏览器自动化等代理任务,性能为迄今最佳。

Life After Switching to Kimi

@0xDevin_ · 6.6K 粉丝 · 38.3K 阅 · 539 赞 · 5 转

Most AI tools are chatbots with a nice interface. Kimi is different. It is a full system: a browser automation engine called Claw that navigates websites like a human, an Agent Swarm that runs

中文介绍 0xDevin_ 分享切换到 Kimi 后的体验:Kimi 包含浏览器自动化引擎 Claw、Agent Swarm 等多智能体协作,是一个完整系统,而非普通聊天界面。

Human in the /loop

@ericzakariasson · 76.0K 粉丝 · 32.3K 阅 · 518 赞 · 29 转

What I like most about coding with agents right now is the room to leave a few runs going and still get on with other work. When something finishes or needs a call, I show up. This post is a short

中文介绍 Eric Zakariasson 谈 agent 编程心得:可同时运行多个 agent 任务,完成或需决策时再介入,保持灵活的人机协作循环。

Previewing GPT-5.6 Sol: a next-generation model

OpenAI previews GPT-5.6 Sol, a next-generation model with stronger capabilities in coding, science, and cybersecurity, paired with its most advanced safety stack.

中文介绍 OpenAI预览下一代模型GPT-5.6 Sol,在编程、科学和网络安全方面能力更强,并配备最先进的安全堆栈。

Repositioning retail for the AI era

Artificial intelligence is rapidly reshaping retail, but not in the ways consumers might immediately notice. The biggest transformation may not be flashy virtual try-ons or chatbot shopping assistants, but in how decisions are made behind the scenes: how products surface in search results, how inven

中文介绍 MIT Tech Review指出AI正重塑零售业,最大变革发生在后台决策,如产品搜索与曝光方式,而非消费者可见的虚拟试穿或聊天机器人。

not much happened today

**Z.ai's GLM-5.2** leads in coding and agent benchmarks with top scores like **1595** on Code Arena: Frontend and **34.29%** reasoning accuracy with zero failures. Databricks improved GLM-5.2 speed to **392 tok/s** using hardware and optimizations. **Ornith-1.0**, a new MIT-licensed coding model fam

中文介绍 Z.ai的GLM-5.2在编码和智能体测试中领先,Code Arena: Frontend得分1595,推理准确率34.29%且零失败。Databricks优化后速度达392 tok/s。新模型Ornith-1.0以MIT许可证发布。

[AINews] It's Meta-Harness Summer

Move over, Harness Engineering, it is time for the harness of harnesses!

中文介绍 Latent Space文章称“Meta-Harness Summer”到来,强调超越传统Harness Engineering,进入更高级的harness阶段。

How agents are transforming work

A new OpenAI research paper shows how AI agents are transforming work, enabling longer, more complex tasks and expanding productivity across roles.

中文介绍 OpenAI新研究论文显示AI智能体正变革工作方式,支持更长更复杂的任务,并提升各岗位生产力。

The emergence of the web data infrastructure layer for AI

AI is booming. New use cases are emerging each day. To capitalize on the technology’s potential, enterprises require data at scale. In many cases, though, the relevant information is blocked or unstructured, which limits its use by AI models. To understand this challenge, consider the foundation of

中文介绍 MIT Tech Review探讨AI时代网页数据基础设施层的兴起,指出企业需要大规模数据,但相关信息常被封锁或非结构化,限制AI模型使用。

市场总览

当前市场技术面整体偏弱,各类资产分化明显。美股方面,标普500 ETF(SPY)和纳斯达克100 ETF(QQQ)近5日分别下跌2.38%和4.60%,RSI均处于40-50中性偏弱区间,且股价跌破20日均线,但标普500仍守在50日均线上方,显示短期回调但中期趋势尚未完全转空。加密资产呈现极度恐慌状态,加密恐慌贪婪指数仅13,总市值2.16万亿美元,比特币(BTC-USD)较52周高点下跌超52%,RSI 31.4逼近超卖,均线空头排列且MACD死叉,以太坊(ETH-USD)同样深陷空头困境。中概股技术面最为疲弱,阿里巴巴(BABA)RSI仅16.5处于严重超卖,股价远低于所有均线,京东(JD)RSI 26.8同样超卖,腾讯(0700.HK)接近52周低点,中概整体呈现连续杀跌后的超卖格局。商品外汇方面,WTI原油(CL=F)RSI 27.9进入超卖区,黄金(GC=F)MACD死叉且绿柱放大;美元指数(DX-Y.NYB)则呈现强势,RSI 71.3超买,多头排列且接近52周高,与多数风险资产形成对比。综合看,市场风险偏好低迷,避险货币和美元走强,超卖资产可能酝酿技术性反弹但趋势尚未反转。

今日关注

DX-Y.NYB 美元指数 (DXY)
偏上行

RSI 71.3处于超买区,价格高于SMA20/50/200形成多头排列,MACD正值且信号线金叉,接近52周高点,短期上行动能充沛,但超买区间需警惕技术性回调。

BABA 阿里巴巴 (BABA)
偏下行

RSI 16.5处于严重超卖区,价格远低于SMA20(113.53)、SMA50(126.3)、SMA200(148.22),呈空头排列,MACD绿柱放大,空头动能强劲,短期超卖后或有反弹但整体下行趋势显著。

BTC-USD 比特币 (BTC-USD)
偏下行

较52周高点下跌52.52%,RSI 31.4接近超卖,价格跌破所有均线(SMA20/50/200),MACD死叉且负值扩大,加密恐慌指数13体现极度恐慌,技术面空头占优。

CL=F WTI 原油期货
偏下行

RSI 27.9进入超卖区,价格低于SMA20(83.29)和SMA50(92.01),且低于SMA200(73.9)显示中长期空头,MACD绿柱持续放大,近5日下跌8.3%,下行压力未减。

SPY 标普500 ETF (SPY)
中性

RSI 43.2处于中性偏弱区间,价格略低于SMA20(743.6)但高于SMA50(734.35)和SMA200(690.54),MACD快线略低于零轴,多空力量均衡,短期方向不明,等待均线进一步指引。

全部资产

^VIX

VIX 恐慌指数

$18.41 -2.54%
5 日
+12.26%
距 52w 高
-47.8%
RSI(14)
51.1
趋势
中性
SMA 20 / 50 / 200
17.92 / 17.79 / 18.64
MACD / 信号
0.152 / 0.038
MACD 金叉 (3 天前)

^TNX

10Y 美债收益率 (%)

$4.37 -1.77%
5 日
-2.56%
距 52w 高
-12.5%
RSI(14)
39.9
趋势
中性
SMA 20 / 50 / 200
4.48 / 4.43 / 4.22
MACD / 信号
-0.005 / 0.012

DX-Y.NYB

美元指数 DXY

$101.37 -0.06%
5 日
+0.51%
距 52w 高
-0.4%
RSI(14)
71.3
趋势
多头
SMA 20 / 50 / 200
100.14 / 99.22 / 98.77
MACD / 信号
0.614 / 0.467
RSI 超买接近 52 周高多头排列

SPY

S&P 500 ETF

$728.99 -0.72%
5 日
-2.38%
距 52w 高
-4.1%
RSI(14)
43.2
趋势
中性
SMA 20 / 50 / 200
743.60 / 734.35 / 690.54
MACD / 信号
-0.138 / 2.657

QQQ

Nasdaq 100 ETF

$706.52 -1.38%
5 日
-4.60%
距 52w 高
-5.6%
RSI(14)
46.5
趋势
多头
SMA 20 / 50 / 200
724.76 / 702.79 / 632.15
MACD / 信号
3.467 / 7.544
多头排列

AAPL

Apple

$283.78 +3.14%
5 日
-4.78%
距 52w 高
-10.6%
RSI(14)
41.3
趋势
中性
SMA 20 / 50 / 200
298.29 / 291.50 / 269.46
MACD / 信号
-2.256 / 0.508

MSFT

Microsoft

$372.97 +5.71%
5 日
-1.69%
距 52w 高
-32.9%
RSI(14)
40.4
趋势
空头
SMA 20 / 50 / 200
400.11 / 410.97 / 447.99
MACD / 信号
-13.872 / -9.812
空头排列

NVDA

Nvidia

$192.53 -1.64%
5 日
-8.62%
距 52w 高
-18.6%
RSI(14)
37.5
趋势
中性
SMA 20 / 50 / 200
207.77 / 210.09 / 190.64
MACD / 信号
-3.697 / -1.793

GOOGL

Alphabet

$337.39 -1.84%
5 日
-8.33%
距 52w 高
-17.4%
RSI(14)
33.3
趋势
中性
SMA 20 / 50 / 200
360.81 / 369.28 / 313.85
MACD / 信号
-7.353 / -4.111

TSLA

Tesla

$379.71 +1.22%
5 日
-5.19%
距 52w 高
-23.9%
RSI(14)
41.2
趋势
空头
SMA 20 / 50 / 200
401.55 / 404.76 / 417.96
MACD / 信号
-7.976 / -4.373
空头排列

META

Meta

$550.25 +1.36%
5 日
-4.67%
距 52w 高
-30.9%
RSI(14)
37.4
趋势
空头
SMA 20 / 50 / 200
583.29 / 612.90 / 650.02
MACD / 信号
-16.259 / -13.176
空头排列
加密恐慌贪婪
13
极度恐慌
加密总市值
$2.16 T
+0.67% / 24h
BTC 主导率
55.7%
ETH 8.8%
24h 成交量
$95.1 B
活跃币 17,449

BTC-USD

Bitcoin

$59,918.95 +0.33%
5 日
-5.25%
距 52w 高
-52.5%
RSI(14)
31.4
趋势
空头
SMA 20 / 50 / 200
63,206.91 / 70,229.14 / 76,000.75
MACD / 信号
-2,296.717 / -2,290.765
MACD 死叉 (今天)空头排列

ETH-USD

Ethereum

$1,573.75 +0.57%
5 日
-7.68%
距 52w 高
-68.2%
RSI(14)
31.7
趋势
空头
SMA 20 / 50 / 200
1,686.24 / 1,920.52 / 2,327.61
MACD / 信号
-77.838 / -78.022
空头排列

SOL-USD

Solana

$71.61 +5.97%
5 日
-1.12%
距 52w 高
-71.7%
RSI(14)
48.9
趋势
空头
SMA 20 / 50 / 200
69.39 / 78.08 / 95.80
MACD / 信号
-1.668 / -2.199
空头排列

BABA

阿里巴巴 (BABA)

$94.81 -0.27%
5 日
-11.48%
距 52w 高
-50.8%
RSI(14)
16.5
趋势
空头
SMA 20 / 50 / 200
113.53 / 126.30 / 148.22
MACD / 信号
-8.528 / -6.804
RSI 超卖空头排列

PDD

拼多多 (PDD)

$76.55 +4.43%
5 日
-3.78%
距 52w 高
-45.1%
RSI(14)
34.2
趋势
空头
SMA 20 / 50 / 200
81.51 / 91.25 / 109.37
MACD / 信号
-4.515 / -4.271
空头排列

JD

京东 (JD)

$25.39 +0.79%
5 日
-7.91%
距 52w 高
-31.1%
RSI(14)
26.8
趋势
空头
SMA 20 / 50 / 200
27.98 / 29.67 / 30.09
MACD / 信号
-1.168 / -0.849
RSI 超卖空头排列

0700.HK

腾讯控股 (0700.HK)

HK$411.80 -2.28%
5 日
-6.45%
距 52w 高
-39.7%
RSI(14)
35.8
趋势
空头
SMA 20 / 50 / 200
445.59 / 461.46 / 561.50
MACD / 信号
-10.926 / -7.749
MACD 死叉 (4 天前)接近 52 周低空头排列

GC=F

黄金期货

$4,103.00 +1.80%
5 日
-2.87%
距 52w 高
-26.6%
RSI(14)
37.9
趋势
中性
SMA 20 / 50 / 200
4,273.07 / 4,488.59 / 4,448.33
MACD / 信号
-119.817 / -106.270
MACD 死叉 (4 天前)

CL=F

WTI 原油期货

$70.24 -2.34%
5 日
-8.30%
距 52w 高
-41.2%
RSI(14)
27.9
趋势
中性
SMA 20 / 50 / 200
83.29 / 92.01 / 73.90
MACD / 信号
-6.453 / -5.185
RSI 超卖

USDCNY=X

美元 / 人民币

¥6.79 -0.00%
5 日
+0.31%
距 52w 高
-5.9%
RSI(14)
56.5
趋势
空头
SMA 20 / 50 / 200
6.77 / 6.79 / 6.95
MACD / 信号
-0.002 / -0.007
接近 52 周低空头排列
风险提示

本报告仅基于历史行情数据的技术指标解读,过去走势不代表未来表现,不构成任何投资建议。仅供技术指标解读参考。

Venezuela Live Updates: Trump’s Vow to ‘Run’ Venezuela Is Tested After Quakes

The U.S. dispatched hundreds of rescue workers to help the Venezuelan government, as the injured overwhelmed hospitals and the death toll rose to 920. The Pentagon sent two ships, transport planes and helicopters.

中文摘要 委内瑞拉地震死亡人数升至920人,美国派遣数百名救援人员及两艘舰船、运输机和直升机协助救援。

Earthquake Tests Growing Ties Between U.S. and Venezuela

The Trump administration said it would commit aid, at a time when it has been expanding U.S. commercial interests in Venezuela beyond oil.

中文摘要 委内瑞拉地震考验美委关系,特朗普政府承诺提供援助,同时正扩大在委内瑞拉的商业利益,不限于石油。

U.S. Strikes Iran in Retaliation for Attack on Vessel in Strait of Hormuz

President Trump on Friday called Iran’s attack on a container ship transiting the Strait of Hormuz a day earlier a “foolish” act.

中文摘要 美国对伊朗发动报复性打击,此前伊朗在霍尔木兹海峡袭击一艘集装箱船,特朗普称袭击为“愚蠢”之举。

US conducts strikes on Iran after attack on cargo ship

US Central Command says it has struck missile and drone storage facilities and coastal radar positions.

中文摘要 美国中央司令部宣布打击伊朗导弹和无人机储存设施及沿海雷达阵地,回应货运船遭袭事件。

What is the framework agreement signed by Israel and Lebanon?

Deal comes after months of US-brokered direct talks, but Israel continues to occupy Lebanese land.

中文摘要 以色列与黎巴嫩在美国斡旋下签署框架协议,但以色列仍占领黎巴嫩土地,协议细节未公开。

Israel and Lebanon sign framework agreement after US-brokered talks

Hezbollah is not party to Friday's agreement, and previous ceasefires between Israel and Lebanon have still seen near-daily cross-border strikes.

中文摘要 以色列与黎巴嫩签署框架协议,真主党未参与。此前停火协议未阻止几乎每日的跨境袭击。

Venezuela quake death toll reaches 920 as interim president vows to save ‘as many as possible’

Delcy Rodríguez says foreign rescue teams are arriving as anger grows at official response and limited resources Survivors tell of ‘brutal and fast’ Venezuela quake as hunt for survivors goes on Venezuela’s interim president, Delcy Rodríguez, has vowed to fight to save “as many people as possible” a

中文摘要 委内瑞拉地震死亡人数达920人,临时总统德尔西·罗德里格斯誓言拯救尽可能多的人,外国救援队陆续抵达,但官方应对迟缓引发不满。

Desperate families search for missing loved ones in Venezuela

Families are sharing details about their loved ones online as more than 50,000 people remain missing in Venezuela.

中文摘要 委内瑞拉地震后超过5万人失踪,绝望的家庭在网上分享亲人信息,试图寻找线索。

Heavyweight champion Usyk says vacating heavyweight world title belts

Usyk, who had held the WBA, WBC and IBF belts, has been the world undisputed cruiserweight and heavyweight champion.

中文摘要 重量级拳王乌西克宣布放弃WBA、WBC和IBF三条金腰带,他曾是次重量级和重量级统一冠军。

In Venezuela’s Rubble, Listening for Whispers And Longing for Help

Facing criticism from residents that it was not doing enough to help, the Venezuelan government said it had dispatched more than 100 heavy machines to clear debris.

中文摘要 委内瑞拉政府面临救援不力批评,称已派出100多台重型机械清理废墟,但居民仍在废墟下等待救援。

US says it struck Iran targets after attack on cargo ship in the strait of Hormuz

Strikes against military facilities were in response to drone attack a day earlier on a cargo vessel US strikes Iran – latest updates The US has struck Iran in a tit-for-tat response to a drone strike on a cargo ship, as the ceasefire between the US and Iran that reopened the strait of Hormuz underg

中文摘要 美国对伊朗军事设施发动打击,报复前一天无人机对霍尔木兹海峡货船的袭击,破坏了两国此前重新开放海峡的停火状态。

Here’s the latest.

中文摘要 美国对伊朗打击事件的最新动态,暂无更多细节。

As Foreign Aid Arrived, Many Venezuelans Asked: ‘Where is Our Military?’

The nation’s armed forces seemed to arrive late into the emergency response — with an emphasis on controlling traffic, even as bodies remained buried under rubble.

中文摘要 委内瑞拉军队在救灾中反应迟缓,重点放在交通管制,废墟下仍有遗体,民众质疑军队未及时救援。

Three unusual things about the King's tax bill

King Charles paid £12.9m in tax for 2024-2025 - here's what we know about his unique tax situation.

中文摘要 英国国王查尔斯在2024-2025财年缴纳了1290万英镑税款,涉及其独特的税务情况。

Earn Your Leisure Co-Founders on Financial Literacy

The CEOs and co-founders of Earn Your Leisure, Rashad Bilal and Troy Millings, explain their educational platform that looks to help individuals gain financial freedom and generational wealth. They discuss how to ride the tech stock wave with Romaine Bostick and Katie Greifeld on Bloomberg's "The Cl

中文摘要 Earn Your Leisure联合创始人Rashad Bilal和Troy Millings介绍其教育平台,旨在帮助个人实现财务自由和世代财富,并讨论如何把握科技股浪潮。

Tennis Pro Caroline Wozniacki on Wimbledon

Tennis champion Caroline Wozniacki sat down with Romaine Bostick and Katie Greifeld on the sidelines of The Wimbledon Court in Central Park. She discusses the future of the sport, player compensation, and advice for her fellow tennis star Serena Williams re-entering competition. (Source: Bloomberg)

中文摘要 网球冠军Caroline Wozniacki接受采访,讨论网球运动的未来、球员薪酬,并就塞雷娜·威廉姆斯复出提出建议。

Hellman & Friedman-Backed Hub Files Confidentially for IPO

Hub International Holdings Inc., an insurance broker backed by Hellman & Friedman, confidentially filed for an initial public offering whose proceeds could be used by the to pare its debt.

中文摘要 由Hellman & Friedman支持的保险经纪公司Hub International Holdings Inc.已秘密提交首次公开募股(IPO)申请,募资可能用于偿还债务。

Invesco’s Kriskey: Oil Price Freefall Is ‘Overdone'

Kathy Kriskey, Invesco's Alterative ETF Strategy Head, helps make sense of oil's drop as ships continue to cross through the Strait of Hormuz despite attacks. She speaks with Romaine Bostick & Katie Greifeld on "The Close." (Source: Bloomberg)

中文摘要 Invesco另类ETF策略主管Kathy Kriskey认为油价暴跌反应过度,指出尽管遭遇袭击,船只仍继续通过霍尔木兹海峡。

Biolife Is Said to Have Drawn Takeover Interest From Repligen

Biolife Solutions Inc. has attracted takeover interest from parties including diagnostics company Repligen Corp., according to people with knowledge of the matter.

中文摘要 据知情人士透露,Biolife Solutions Inc.已吸引包括诊断公司Repligen Corp.在内的多方收购兴趣。

Markets, Economy Ignoring Policy Shocks: Skelly

Dan Skelly, managing director and head of the equity model portfolio team at Morgan Stanley Wealth Management, joins Lisa Mateo and Tom Keene on "Bloomberg Money." (Source: Bloomberg)

中文摘要 摩根士丹利财富管理股票模型投资组合团队主管Dan Skelly表示,市场和经济正在忽视政策冲击。

Dow Jones Finishes Day Down; Week Up | Closing Bell

Comprehensive cross-platform coverage of the U.S. market close on Bloomberg Television, Bloomberg Radio, and YouTube with Romaine Bostick, Katie Greifeld, Carol Massar and Tim Stenovec. (Source: Bloomberg)

中文摘要 道琼斯指数当日收跌,但本周累计上涨。彭博多平台报道美国市场收盘情况。

Coffee Company Kahawa 1893 Works to Empower Women Farmers

Margaret Nyambumo, founder of coffee company Kahawa 1893 and third-generation coffee farmer, said that she created the option to directly tip the female farmers who grow her coffee beans because she wanted to find a way to consistently and fairly compensate them regardless of fluctuation in coffee p

中文摘要 咖啡公司Kahawa 1893创始人Margaret Nyambumo(第三代咖啡农)设立直接给女性种植户小费的选项,旨在公平补偿她们。

Education 'Getting Act Together' on Personal Finance: Cohen

Abby Joseph Cohen, professor at Columbia Business School, joins Lisa Mateo and Tom Keene on "Bloomberg Money." They discuss personal finance and the US education system's effort to educate students on financial literacy. (Source: Bloomberg)

中文摘要 哥伦比亚商学院教授Abby Joseph Cohen讨论个人理财及美国教育系统加强金融知识教育的努力。

Trump threatens 100% tariff on European nations over tech tax

The US president says "Numerous European countries" have been discussing bringing in such a levy.

中文摘要 美国总统特朗普威胁对欧洲国家征收100%关税,原因是这些国家讨论征收数字技术税。

Alternative Soda Brands Taking Over Beverage Markets

Poppi Co-Founder Allison Ellsworth describes struggles of breaking into beverage industry, warns getting product on shelves is "capital intensive." Ellsworth says movement within "MAHA" is changing industry standards, including for beverage giants like Pepsi which purchased Poppi in 2025 for nearly

中文摘要 Poppi联合创始人Allison Ellsworth表示进入饮料行业困难,上架需要大量资金;MAHA运动正改变行业标准,百事可乐等巨头已收购Poppi。

【翰林文苑】基准已发,一键检查你的Opus 4.6是不是真的!

本帖使用社区开源推广,符合推广要求。我申明并遵循社区要求的以下内容: 我的帖子已经打上 开源推广 标签: 是 我的开源项目完整开源,无未开源部分: 是 我的开源项目已链接认可 LINUX DO 社区: 是 我帖子内的项目介绍,AI生成、润色内容部分已截图发出: 是 以上选择我承诺是永久有效的,接受社区和佬友监督: 是 以下为项目介绍正文内容,AI生成、润色内容已使用截图方式发出 一个基于概率分布识别任意模型真假的项目,从此告别掺假! - 开发调优 - LINUX DO github.com GitHub - hanlinwenyuan/hlwy-ai-checker: 检查第三方AI API是

我终于知道deepseek为什么这么急于开发自己的agent了

大模型能力是引擎,agent才是完整的车,引擎固然重要,但是照现在的趋势,引擎之间的差距会越来越小,后面就是看各家的工具对自己模型的适配度,这才是后续的生态护城河,例如claudecode和codex。 我最近在读opencode和pi的源码,越来越能体会到一套好的harness能发挥出多大的作用了。比如我现在在开发一个类silly tavern的桌面端agent,由于原生的并不好用,太多技术债了,我就用pi的agent二开了一套,可以兼容导入社区的预设、角色卡。设计好一套harness,规范ai剧情演绎、设计一系列tool来强制ai进行记忆,有效防止长文本聊天下的记忆丢失和风格偏移问题,并且

【打破信息茧房】说出你爱看的油管博主

最近刷油管被算法焊死在舒适区了,越刷内容越重复。开个楼大家互相安利私藏博主,一起跳出信息茧房! 先抛块砖: 零度解说:挖尽神器,无脑跟随操作 李永乐老师:黑板科普,数理通解 試當真:搞笑短片、网络文化、社会讽刺为主 ,下饭片,10分钟一集 发帖前请过目: 本贴纯属内容交流,不涉及任何政治话题,请大伙自觉遵守站规,聊博主、聊内容、聊干货; 欢迎跟贴分享你私藏的宝藏博主(最好附上主页链接),方便各位佬查阅; 我这点推荐只是引子,希望大家一起把楼盖起来,真正打破信息茧房(AI相关的就更好噜)。 感谢佬们! 有好看的记得喊我,我先去蹲更新了 51 个帖子 - 43 位参与者 阅读完整话题

分享一下我是怎么在一晚上赚¥11000的

涨价前在淘宝 ¥38769 购入,现价 ¥49889 2026款-16英寸 M5 Max 芯片(18+40) 2T固态硬盘 128GB 遵循佬友的意见改了标题,从一个月内变成一晚上 21 个帖子 - 16 位参与者 阅读完整话题

【九幺】恢复了+一些事务

本帖使用社区公益推广,符合推广要求。我申明并遵循社区要求的以下内容: 我的项目是免费使用的,无收费(变相收费、赞助)部分: 是 我的帖子已经打上 公益推广 标签: 是 我的项目属于个人项目,与公司或商业机构无关: 是 我的项目不存在QQ、TG等群组引流: 是 我的项目不存在非运营必要的网站引流: 是 我的项目不存在为他人推广、AFF: 是 我的项目无关联的商业项目: 是 我的站点存在登录,并已接入 LINUX DO Connect: 是 我帖子内的项目介绍,AI生成、润色内容部分已截图发出: 是 以上选择我承诺是永久有效的,接受社区和佬友监督: 是 以下为项目介绍正文内容,AI生成、润色内容已

L 站有人对 Claude Code 泄露的源码感兴趣的吗?

我打算重写一下 Claude Code 技术架构拆解的系列文章。 基于 Claude Code 泄露的源码进行分析。 说打算写,其实我已经正在写了。 我之前写过一个 Claude Code 的架构图,那篇文章我现在看来,还是差了很多东西。 这个课题在我的脑海里有一阵子了。 我网上找到了一些源码分析的文章。 但是绝大多数都是 AI 味儿很大的那种。 要么逻辑衔接不太顺畅,要么虎头蛇尾的。拼接感太严重。 不过有一些作图做的确实比较好,这个我承认。 而且看图并不能让你完整的理解,只能提高一下你对技术图的审美。 还是需要对图进行解释和描述。 不过呢,我已经加上了 gif 图,而且还在尝试加入一些可交互

外区APP Store软件更新技巧

把外区 Apple ID 添加到“设置-备忘录、邮件、日历”里,三个里面随便选一个,之后更新软件可以直接更新。不用去 App Store 来回切账号。 46 个帖子 - 45 位参与者 阅读完整话题

两周年快乐!!顺便来推广一下我的最美cli! 🫣

本帖使用社区开源推广,符合推广要求。我申明并遵循社区要求的以下内容: 我的帖子已经打上 开源推广 标签: 是 我的开源项目完整开源,无未开源部分: 是 我的开源项目已链接认可 LINUX DO 社区: 是 我帖子内的项目介绍,AI生成、润色内容部分已截图发出: 是 以上选择我承诺是永久有效的,接受社区和佬友监督: 是 以下为项目介绍正文内容,AI生成、润色内容已使用截图方式发出 首先恭喜自己,今天注意到头像附近有小蛋糕了,才发现不知不觉加入社区已经两年了 这两年在L站学到了很多,也见证了很多,感谢各位佬友,感谢始皇的付出! 最近有在做一个Cli聚合管理的小软件 【开源自荐】Cli-Manage

不多说,快抢注agent邮箱去

现在可以自定义了,还有比较多的好号 Agent Mail Agent Mail — QQ 邮箱为 Agent 打造的原生邮箱 与个人邮箱隔离,原生适配 Agent,让 AI 智能体安全高效地帮你收发、整理邮件。 269 个帖子 - 245 位参与者 阅读完整话题