#ce
共 558 篇相关内容 · 安全漏洞、AI动态、技术文章
技术文章 159
PHP2Shell 与 Struts 四连补:PHP / Java 生态十月高危处置清单
PHP 10-21 紧急修复 CVE-2026-109618/109619(9.2分 RCE,在野利用),Struts 10-05 四连修含 OGNL 注入 RCE;附版本表、backport 路径与 Oracle 十月 CPU 排期。
GitHub 周报:DeepSeek Harness、OpenViking 与 Agent 生态爆发
本周 GitHub 热点被 AI Agent 基础设施占领:DeepSeek Harness 14 万星破纪录、OpenViking 统一上下文层、Prime Agent、career-ops、ai-memory 等集体上榜。一文速览值得关注的项目。
开源权重模型 vs 闭源 API:企业 AI 选型的分水岭
硅谷数据显示企业推理 token 支出正从闭源流向开源权重模型,DoorDash、Airbnb、Coinbase 等已转用中国开源模型。安全与合规考量成为选型关键变量,本文给出权衡框架。
CVE-2026-73678:MindsDB 未授权 RCE 敲响 AI 基础设施安全警钟
MindsDB 的 Minds Platform 存在 CVSS 10.0 的未授权远程代码执行漏洞,攻击者可无需认证在服务器执行任意命令。AI 数据基础设施成为高危目标,本文给出排查与修复建议。
CVE-2026-74938:Firefox 垃圾回收缓解绕过漏洞解析
Firefox 与 Thunderbird 的 JavaScript 垃圾回收缓解机制存在绕过漏洞,CVSS 9.8,攻击者可通过特制网页触发内存破坏实现远程代码执行。建议尽快升级。
Azure Service Bus 9.9 分漏洞与 8 月云安全要点
8 月补丁日微软修复 415 个漏洞,其中 Azure Service Bus RCE(CVSS 9.9)、Azure Confidential Ledger RCE(9.1)、Teams RCE(10.0)等云服务漏洞值得云架构团队优先关注。
vm2 沙箱逃逸 (GHSA-cfcw-xp6x-25gj):维护已停,迁移要趁早
Node.js 沙箱库 vm2 曝出 CVSS 9.8 的沙箱逃逸漏洞,可逃逸到宿主机执行任意代码。项目已停止维护,本文解析漏洞背景并给出迁移方案对比。
供应链攻击的下一个目标:Cursor 与 Claude Desktop 生态
Intel 471 报告指出,攻击者正把投毒目标从传统 npm/PyPI 转向 AI 编程助手生态:恶意 skill、插件与"流行包代理"手法瞄准 Cursor、Claude Desktop 等 Agent 环境。本文解析攻击面与防护建议。
Prime Agent 与递归语言模型:把 Agent 上下文变成代码
Prime Intellect 开源的 Prime Agent 单日 2319 stars 登顶 GitHub Trending。它让模型持有持久 Python 会话,把文件、命令与子 Agent 全当作代码管理,基于 arXiv:2512.24601 的递归语言模型思想。
OpenViking:字节跳动开源的统一上下文层,Agent 记忆的新范式
字节跳动火山引擎开源 OpenViking,把 Agent 记忆、知识 RAG 与技能合并为单一自进化上下文层,29K stars 单日新增 213。本文拆解其设计思路、适用场景与社区插件生态。
DeepSeek Harness:14 万星的开源 Agent 运行时到底做了什么
DeepSeek 开源 Agent 运行时 DeepSeek Harness 一小时内突破 2 万 stars,四天超 13 万,创 GitHub 史上最快涨星纪录。它用"一切皆插件"的设计统一多种模型平台,本文拆解其架构、执行模式与生态价值。
2026 年 7 月勒索软件态势:新家族崛起,AI 智能体攻击链登场
360 发布 2026 年 7 月勒索软件流行态势分析:ExfilSquad、CRPxO、GSG、Section9、Gammax 五大新家族登场,BrzCrypt 与 Wmansvcs 判定同一团伙。同期 Sysdig 披露首个 AI 智能体勒索攻击链,CVERC 预警 Sorry 勒索软件攻击 Linux 服务器。
GitHub 宕机 7 小时背后:Agent 时代的代码托管之争,Cursor Origin 想颠覆什么
2026 年 8 月 17 日 GitHub 发生近 7 小时大规模宕机,同一天 Cursor 开放 Agent 原生代码托管平台 Origin 测试。本文复盘事件始末,拆解 Origin 的合并队列、MCP 支持与镜像加一键 Detach 的寄生式进攻策略,讨论 Agent 时代代码托管基础设施的变革。
8月第四周 AI 与开源动态:OpenViking 上下文层与生态展望
字节跳动开源 OpenViking 统一上下文层引发关注。本文展望 Agent 上下文层赛道与 8 月开源生态走向。
CVE-2026-74938:Firefox GC 缓解绕过技术解析
CVE-2026-74938 是 Firefox/Thunderbird 的 JavaScript 垃圾回收缓解绕过漏洞(CVSS 9.8)。本文解析机制、风险与修复。
vm2 沙箱逃逸(GHSA-cfcw-xp6x-25gj)迁移实战
vm2 曝出 CVSS 9.8 沙箱逃逸漏洞且已停止维护。本文给出从 vm2 迁移到替代方案的实战步骤与代码示例。
GitHub 周报:career-ops 与 AI 求职自动化
AI 求职工具 career-ops 以 6.6 万星上榜。本文盘点 AI 求职与垂直 Agent 工具的开源趋势。
企业 AI 模型选型:开源权重 vs 闭源 API 实操框架
企业推理 token 支出正向开源权重模型迁移,但安全与合规顾虑并存。本文给出一套企业模型选型的实操权衡框架。
DeepSeek Harness 插件开发:Agent 生态的下一站
DeepSeek Harness 以插件化架构破纪录涨星。本文解析 Harness 的插件模型与开发者如何在其中构建生态。
JadePuffer:首个 AI 智能体勒索攻击链拆解
Sysdig 披露 JadePuffer 活动:攻击者利用 AI 智能体自动化侦察、横向移动与勒索部署,被认为是首个以 AI 智能体为核心编排的勒索攻击链。
360 发布2026年7月勒索软件态势:五大新家族与同团伙判定
360 发布 7 月勒索软件流行态势分析:ExfilSquad、CRPxO、GSG、Section9、Gammax 五大新家族登场,BrzCrypt 与 Wmansvcs 判定同一团伙。
Azure 云安全:Service Bus 与 Confidential Ledger 漏洞处置
Azure Service Bus 与 Confidential Ledger 相关漏洞引发关注。本文解读云服务漏洞的处置逻辑与租户侧审计要点。
2026年8月安全前瞻:勒索、供应链与 AI 攻击三条主线
展望 8 月安全态势:勒索软件新家族、供应链投毒扩散至 AI 生态、AI 基础设施攻击。本文给出三条主线的防御优先级。
reverse-skill 登顶 Trending:攻防 Agent 技能路由器
reverse-skill 登上 GitHub Trending 榜首:面向逆向与渗透测试的 Agent 技能路由器,支持 Claude Code、Codex、Cursor。本文解析其设计与安全价值。
GPT-5.6 Luna 降价 80%:AI 推理成本下行的连锁反应
OpenAI 将 GPT-5.6 Luna 价格下调 80%、Terra 下调 20%。本文解读降价动因、对开发者与 Agent 应用的影响。
GitHub 周报:DeepSeek Harness 之前,Agent 运行时演化史
在 DeepSeek Harness 爆火之前,Agent 运行时经历了从框架到基础设施的演化。本文梳理 Agent 运行时的发展脉络与未来方向。
OpenAI 硬性支出上限全量上线:API 成本治理新常态
7月23日起 OpenAI API 平台为所有账户提供硬性月度支出上限,超限返回 429。本文解读功能边界与成本治理建议。
CVE-2026-62815:Microsoft QUIC 协议栈 RCE 技术解析
微软 QUIC 协议栈被披露 CVSS 9.8 的远程代码执行漏洞 CVE-2026-62815。本文解析 QUIC 攻击面、利用场景与修复优先级。
GitHub 周报:Agent 技能、记忆与上下文层的开源热潮
本周 GitHub Trending 被 Agent 基础设施占领:技能合集、记忆库与上下文层项目扎堆上榜。本文盘点值得关注的仓库与社区信号。
CVE-2026-73939:Oracle Helidon 认证绕过漏洞深度解析
Oracle Helidon 被披露 CVSS 8.6 的认证绕过漏洞 CVE-2026-73939,攻击者可无认证访问受保护资源。本文解析漏洞成因、影响范围与修复方案。
2026年7月第二周安全周报:零日、勒索与AI攻击面速览
7月第二周安全动态:Windows 组件漏洞、AI 基础设施攻击面与勒索软件家族演变。一文汇总本周值得关注的安全信号与处置优先级。
大脑GPS的数学密码:拓扑学如何破解网格细胞的空间导航之谜
# 大脑GPS的数学密码:拓扑学如何破解网格细胞的空间导航之谜 > 你闭上眼睛,在黑暗中走过一条熟悉的走廊——你依然知道自己的位置。这种神奇的空间感知能力,源于大脑深处一组被称为"网格细胞"的神经元。最新研究表明,数学中的拓扑学可以帮助我
2026年AI自动化选品实战教程:Python+SQLite+机器学习从零搭建电商爆品预测系统,含完整代码
# 第3课:AI自动化选品 > 学完这节课,你将能够用Hermes Agent和Python实现全自动选品,从海量商品中筛选出最值得推广的爆品。 --- ## 一、选品逻辑详解 ### 1.1 选品的核心指标 选品不是随便找个高佣
Cisco CUCM漏洞披露后24小时内即被武器化:企业通信系统面临零日威胁
Cisco CUCM关键RCE漏洞在公开披露后不到24小时就被攻击者武器化利用。CVSS 9.8分,无需认证即可远程执行代码。企业通信基础设施正成为攻击者的重点目标。
CVE-2026-31431 Copy Fail:732字节Python脚本攻破Linux内核,容器逃逸零门槛
Linux内核AF_ALG加密子系统存在潜伏9年的逻辑漏洞,732字节Python脚本即可实现容器逃逸和root提权。同一份exploit无需修改可在Ubuntu、RHEL、SUSE、Amazon Linux上通用运行。Kubernetes默认seccomp配置无法阻止利用。
CVE-2026-0300 Palo Alto PAN-OS Zero-Day: Unauthenticated Root RCE Under Active Exploitation
Critical buffer overflow in Palo Alto PAN-OS Captive Portal (CVE-2026-0300, CVSS 9.3) enables unauthenticated root RCE. Nation-state group CL-STA-1132 actively exploiting. CISA KEV listed. Patch immediately.
CVE-2026-2441 Chrome 2026 First Zero-Day: CSS UAF Allows Sandboxed RCE
Google patches first actively exploited Chrome zero-day of 2026. CVE-2026-2441 is a CVSS 8.8 use-after-free vulnerability in the CSS engine that allows remote code execution via crafted HTML pages. Over 3 billion Chrome users at risk.
2026年最值得关注的7款开发者新工具深度解析
AI Agent、CLI革命、开源浪潮——深度解析Antigravity CLI、OpenCode、Claude Code、Cursor、即梦AI CLI、Augment Code、moonrepo 7款2026年最新发布的重磅开发者工具,含安装教程、代码示例和趋势分析。
CVE-2026-42897:Microsoft Exchange Server零日漏洞——打开邮件即可被攻击
微软2026年6月Patch Tuesday修复Exchange Server零日漏洞CVE-2026-42897,CVSS 8.1。攻击者通过恶意邮件实现XSS/CSS注入,Pwn2Own Berlin比赛中被展示完整利用链。
CVE-2026-34621:Adobe Reader零日漏洞深度分析——恶意PDF即可远程执行代码
Adobe Reader零日漏洞CVE-2026-34621从2025年12月起在野外被利用,攻击者通过恶意PDF文件实现远程代码执行。CVSS 8.6,影响所有Adobe Acrobat/Reader版本。
2026年最值得关注的6款开发者AI工具:从终端到IDE的全面革新
深度解析2026年6款最新AI开发工具:Antigravity CLI、GitHub Copilot CLI、OpenAI Codex CLI、OpenCode、Aider、Visual Studio 2026,含功能对比表、代码示例和安装指南
CVE-2026-42530与CVE-2026-42055:NGINX双Critical漏洞可实现远程代码执行
F5发布带外安全更新修复NGINX两个CVSS 9.2严重漏洞,攻击者可利用HTTP/3和HTTP/2模块实现远程代码执行,影响数百万Web服务器。
CVE-2026-7482 Bleeding Llama:Ollama严重内存泄露漏洞,30万台AI服务器裸奔在公网
Ollama曝出CVSS 9.1严重漏洞,攻击者仅需3次API调用即可泄露服务器全部进程内存。全球约30万台暴露在公网的Ollama实例面临即刻威胁。
Firefox Thunderbird CVE-2026-12293 WebGPU Use-After-Free CVSS 9.8 Remote Code Execution
Mozilla Firefox and Thunderbird patch CVSS 9.8 WebGPU Use-After-Free vulnerability allowing remote code execution with zero user interaction. Fixed in version 152. Affects all platforms.
Splunk Enterprise CVE-2026-20253 Pre-Auth RCE CVSS 9.8 PostgreSQL Sidecar Attack
Splunk Enterprise 10.x exposes CVSS 9.8 pre-auth RCE via PostgreSQL sidecar service endpoint with no authentication. CISA KEV listed June 18. Affects 10.0-10.2.3, fixed in 10.2.4/10.0.7.
Check Point VPN Zero-Day CVE-2026-50751: CVSS 9.3 IKEv1 Auth Bypass, Qilin Ransomware Exploited for 31 Days
CVE-2026-50751 is a CVSS 9.3 authentication bypass in deprecated IKEv1 VPN protocol. Exploited from May 7 to June 8, 2026 before Check Point released a patch. CISA escalated to KEV in under 3 hours. Qilin ransomware affiliates leveraged it for initial access.
CVE-2026-7482 Bleeding Llama: Ollama Critical Memory Leak Affects 300K AI Servers
CVE-2026-7482 (CVSS 9.1) in Ollama allows unauthenticated attackers to dump entire process memory via 3 API calls, leaking API keys, system prompts, and credentials from 300K+ exposed servers.
Spring for GraphQL CVE-2026-41699: Paginated Query Deserialization RCE Analysis
Spring for GraphQL框架在处理分页Connection类型查询时存在不安全反序列化漏洞(CVSS 8.1),攻击者可构造恶意GraphQL请求实现远程代码执行,影响1.3.0至2.0.3全版本。
MariaDB CVE-2026-49261 CVSS 10.0 RCE: wsrep_notify_cmd Shell Injection Triple Threat
MariaDB连续披露三个严重漏洞:CVE-2026-49261(CVSS 10.0满分RCE)、CVE-2026-44170(Windows命令注入)和CVE-2026-44172(big5编码SQL注入绕过),影响10.6.x到12.3.x全线版本。
The Gentlemen勒索软件深度分析:90/10分成模式如何催生2026年第二活跃的RaaS组织
The Gentlemen勒索软件组织以90/10利润分成模式迅速崛起为2026年第二活跃的勒索软件团伙,已公布332名受害者。Check Point揭露运营者为来自俄罗斯伊热夫斯克的黑客。
Ingram Micro勒索软件泄露事件:全球IT分销巨头42,000名员工数据遭Safepay组织窃取
全球最大IT分销商Ingram Micro遭Safepay勒索软件攻击,42,521名员工SSN和财务数据被窃取。从攻击到披露耗时6个月,暴露供应链巨头在事件响应中的结构性问题。
Kubernetes CVE-2026-1483容器逃逸漏洞:从Pod到宿主机Root的完整攻击链
CVE-2026-1483允许拥有Pod创建权限的攻击者突破容器隔离获取宿主机root权限。本文详解攻击链、同类逃逸漏洞对比、Falco检测规则和三层防御加固方案。
Ivanti Sentry CVSS 10.0 RCE Vulnerability CVE-2026-10520 Deep Analysis
CVE-2026-10520 (CVSS 10.0) allows unauthenticated root RCE in Ivanti Sentry. watchTowr published PoC on June 10, two devices confirmed backdoored same day. CISA added to KEV catalog. Full technical analysis, IoC detection, and enterprise response checklist.
TeamPCP供应链攻击170+个npm PyPI包被同时投毒
TeamPCP compromised 170+ npm and PyPI packages targeting TanStack, Mistral AI and OpenSearch ecosystems in coordinated supply chain attack.
Murray County支付20万美元赎金:美国县级政府为何成为勒索软件的提款机
乔治亚州Murray County向勒索软件攻击者支付20万美元赎金,LockBit 4.0变种加密税务和法院系统。县级政府的网络安全困境与务实防御路线。
Ollama Bleeding Llama漏洞(CVE-2026-7482):30万台AI推理服务器面临内存泄露风险
Cyera发现Ollama严重未认证内存泄露漏洞CVE-2026-7482(CVSS 9.1),攻击者无需凭据即可读取进程内存。全球约30万台暴露实例受影响,附检测脚本和修复方案。
用Python复刻Shodan:4000行代码自建网络空间搜索引擎
# 用Python复刻Shodan:4000行代码自建网络空间搜索引擎 > 当你在Shodan上搜索"Nginx country:CN"时,背后发生了什么?端口扫描、Banner抓取、指纹匹配、地理定位、数据索引——这些核心能力,能否用纯
用Python复刻Nessus:个人主机漏洞扫描器完整实现
## 前言 在企业安全建设中,Nessus 作为商业漏洞扫描器的标杆,几乎是安全团队的标配工具。然而其高昂的商业授权费用让很多个人开发者和小型团队望而却步。本文将手把手带你用 Python 从零实现一个功能完备的个人主机漏洞扫描器——**
用Python复刻Burp Suite核心功能:个人Web安全扫描器完整实现
# 用Python复刻Burp Suite核心功能:个人Web安全扫描器完整实现 ## 引言 Burp Suite是Web安全领域最知名的渗透测试工具,其企业版价格不菲,社区版功能受限。本文将介绍如何用纯Python实现一个具备Burp
开源AI视频生成:CogVideoX/Open-Sora替代Runway的现状与展望
# 开源AI视频生成:CogVideoX/Open-Sora替代Runway的现状与展望 AI视频生成是2024-2025年最火的赛道。Runway Gen-3月费$15起步,Pika也要$10/月,而且生成时长、次数都有限制。开源社区有
本地AI绘画替代Midjourney:Stable Diffusion/FLUX完全指南
# 本地AI绘画替代Midjourney:Stable Diffusion/FLUX完全指南 每月花$10-$60订阅Midjourney?DALL-E 3按token烧钱?如果你有一张显卡,完全可以把这些钱省下来,用开源工具在本地实现同
免费部署方案:Cloudflare Pages+Coolify替代Vercel/Netlify,一年省下3000元
## 前言:你的部署平台账单有多高? 作为一名独立开发者或小团队,每个月在Vercel Pro($20/人/月)或Netlify Pro($19/人/月)上的开销看似不多,但一年下来每个用户就是240美元(约1700元人民币)。如果你是5
ZAP+Nuclei+Nikto:开源DAST工具链最佳组合
# ZAP+Nuclei+Nikto:开源DAST工具链最佳组合 ## 一、DAST是什么?为什么你需要它? 动态应用安全测试(DAST)是在应用程序运行状态下进行黑盒安全扫描的技术。与SAST(静态分析)不同,DAST不需要源代码,直
Shuffle+TheHive:开源SOAR安全编排自动化实战
# Shuffle+TheHive:开源SOAR安全编排自动化实战 ## 一、为什么需要SOAR? 在现代安全运营中心(SOC)中,安全团队每天面对海量告警,手动处理效率极低。安全编排、自动化与响应(SOAR)平台能够将安全工具串联成自
SpiderFoot+Amass:免费OSINT自动化情报收集平台搭建
# SpiderFoot+Amass:免费OSINT自动化情报收集平台搭建 ## 前言 在渗透测试和安全评估中,信息收集是至关重要的第一步。Maltego Pro作为商业OSINT工具的标杆,月费$99起,企业版更是高达$9,999+/
Sliver vs Cobalt Strike:2025开源C2框架深度对比
# Sliver vs Cobalt Strike:2025开源C2框架深度对比 > 本文面向红队工程师和安全研究员,从实战角度对比主流C2框架,提供完整安装部署指南。 ## 一、为什么关注开源C2框架? Cobalt Strike一
MISP+OpenCTI:构建企业级开源威胁情报平台
## 前言:威胁情报不是玄学 在安全运营中心(SOC)的日常工作中,分析师经常面对海量告警——"这个IP是不是恶意的?"、"这个域名关联了哪个APT组织?"、"这个hash是什么时候被发现的?"。没有威胁情报平台(TIP),这些问题只能靠
Keycloak+Headscale:搭建开源零信任网络访问方案
## 前言:为什么需要零信任? 传统网络安全模型基于"内网可信、外网不可信"的边界思维,但随着远程办公、云原生架构和BYOD的普及,这种模型已经千疮百孔。零信任网络访问(ZTNA)的核心理念是"永不信任,始终验证"——无论用户身处公司内网
ModSecurity+Coraza:开源WAF部署与规则编写指南
# ModSecurity+Coraza:开源WAF部署与规则编写指南 ## 为什么商业WAF不是唯一选择? 企业WAF市场被Imperva、F5、Cloudflare三大厂商瓜分,年费动辄上万美元。对于中小企业和独立开发者来说,这笔开
自托管Git服务选型:Gitea vs Forgejo vs GitLab CE全面对比
# 自托管Git服务选型:Gitea vs Forgejo vs GitLab CE全面对比 ## 为什么你需要逃离付费Git托管? GitHub Team每用户每月$4,GitLab Premium直接飙到$29/月/人。一个10人团
Semgrep+CodeQL替代Checkmarx:开源SAST最佳实践
# Semgrep+CodeQL替代Checkmarx:开源SAST最佳实践 ## 为什么我们要逃离商业SAST? 每年花5万美金买Checkmarx授权,结果扫描出来的漏洞还是靠人工review确认——这是很多安全团队的真实写照。商业
Wazuh+Graylog替代Splunk:零成本SIEM方案搭建实战
# Wazuh+Graylog替代Splunk:零成本SIEM方案搭建实战 企业级SIEM(安全信息与事件管理)是安全运营的核心,但商业方案的价格令人望而却步。Splunk Enterprise起步价$1,800/GB/年,最低$15,0
开源设计工具Penpot和Excalidraw:Figma的免费自托管替代方案
# 开源设计工具Penpot和Excalidraw:Figma的免费自托管替代方案 ## 为什么设计师需要免费替代方案? Figma是目前最流行的UI设计工具,但它并不便宜。每个编辑者每月$15,一个5人设计团队一年就要$900。更让人
Postman免费替代方案:Bruno、Hoppscotch、HTTPie深度对比
# Postman免费替代方案:Bruno、Hoppscotch、HTTPie深度对比 ## 为什么你该抛弃Postman? Postman曾经是每个开发者的必备工具,但现在它已经变成了一个臃肿的商业平台。免费版限制越来越多:集合数量受
Burp Suite Pro免费替代:用ZAP+Nuclei搭建个人Web安全测试平台
# Burp Suite Pro免费替代:用ZAP+Nuclei搭建个人Web安全测试平台 ## 前言 Burp Suite Pro是Web安全测试领域的"瑞士军刀",但每年$449的许可证费用让很多独立安全研究员和中小企业望而却步。本
DeepSWIP:让神经符号系统学会反事实推理
# DeepSWIP:让神经符号系统学会反事实推理 ## TL;DR 神经符号系统(如DeepProbLog)结合了神经网络的感知能力和逻辑推理的精确性,但它们的推理方式是"关联性"的——只能回答"给定证据,什么最可能发生?"。这篇论文
LedgerAgent:用结构化状态让AI客服代理遵守规则
# LedgerAgent:用结构化状态让AI客服代理遵守规则 ## TL;DR 在客服场景中,AI代理需要跨多轮对话维护任务状态、调用工具并遵守业务策略。现有的代理架构把所有信息混在提示词中,让模型自己"重建"状态,导致策略违规和状态
Wazuh/Graylog免费替代方案:本地搭建企业级SIEM安全监控平台全指南
Splunk 是SIEM领域的霸主,但其定价模型堪称"恐怖"——按每日数据摄入量(GB/天)计费,500GB/天的环境年费轻松突破 $100,000。Wazuh(开源XDR/SIEM)+ Graylog(日志管理)的组合可以完全替代Splu
当量子实验室没有固定因果序:GPU加速的半定规划如何叩问因果博弈的极限
Boghiu和Simonov利用GPU加速的半定规划求解器,系统探索局域维度d=5至8下GYNI博弈的最优过程矩阵策略,发现提高维度无法改善已知最佳获胜概率0.6218,暗示需要本质不同的策略或更紧的上界。
四倍速基因组比对:minibwa如何重新定义短读长序列映射的性能边界
Heng Li与Nils Homer联合发布minibwa,在保持BWA-MEM级精度的前提下实现四倍加速,融合minimap2链式比对与智能预取策略,为大规模基因组分析流水线带来实质性性能突破。
2026年AI推理基础设施大战:Baseten 15亿美元融资背后的产业逻辑
Baseten五个月估值从50亿飙至130亿美元,Cognition AI以250亿美元估值融资10亿,推理赛道正在成为AI产业的新战场。
用对齐而非复杂度来解码大脑:线性对比解码器如何胜过深度网络
认知科学认为大脑用高维向量空间表征概念。本文发现,从fMRI信号解码视觉、语言、音频刺激时,线性对比解码器一致优于岭回归和非线性深度网络。关键洞见:解码性能提升来自对比学习目标的选择,而非模型架构复杂度。fMRI的空间-时间平均效应将非线性神经计算线性化,使得简单线性模型配合适当目标函数即可达到最优。
单个神经元能'说话'吗?NEURRATOR用自然语言解读视觉皮层的细胞级语义
本文解读NEURRATOR框架——首个将单个神经元的放电活动解码为自由形式自然语言描述的系统。该框架结合CLIP编码器与多模态语言模型,在小鼠视觉皮层Neuropixels记录上实现了细胞级别的场景叙述,让每个神经元都能用人类语言'讲述'它看到了什么。
ev-flow:基于NHTS数据的美国八区域电动汽车充电行为合成生成器
ev-flow是一个开源Python工具包,基于美国2017年全国家庭出行调查(NHTS)数据,通过九阶段确定性流水线为美国八个地区生成行为真实的电动汽车充电合成数据。工具采用三层伯努利插电模型、K=16高斯混合时间采样和温度依赖能耗修正,在湾区住宅验证中21项检查零未解释失败,填补了美国本土EV充电数据生成的空白。
ev-flow:基于NHTS的美国八区域电动汽车充电行为合成数据生成器
ev-flow是一个开源Python工具包,基于美国2017年全国家庭出行调查(NHTS)数据,通过九阶段确定性流水线为美国八个地区生成行为真实的电动汽车充电合成数据。工具采用三层伯努利插电模型、温度依赖能耗修正和区域销售混合模型,在湾区住宅场景验证中21项检查零未解释失败,填补了美国本土EV充电数据生成的空白。
量子动力学模拟的突破:保守自适应秩方法求解Wigner-Poisson系统
密歇根州立大学团队提出一种针对Wigner-Poisson系统的保守自适应秩方法,通过费米-狄拉克型重构和全局能量约束校正,在保持物理保真度的前提下大幅降低量子动力学模拟的计算成本。该方法在双流不稳定性、强朗道阻尼等经典基准测试中展现出接近机器精度的守恒误差。
Atomic Arch:400+个Linux AUR包被劫持植入Rust窃密器和eBPF Rootkit,开发者信任模型崩塌
2026年6月Arch Linux AUR遭遇史上最大供应链攻击,400+孤儿包被接管植入Rust窃密器和eBPF内核rootkit。深度分析PKGBUILD注入机制、eBPF隐匿技术、应急排查和Linux生态安全治理方案。
等离子体波导中的无失相激光尾场加速:突破电子加速的物理极限
Palastro等人提出在等离子体波导中叠加多种模式构造时空结构化激光脉冲,以真空中光速驱动尾场,从根本上消除了激光尾场加速器的失相限制。方案保持恒定光斑尺寸和超短脉冲持续时间,大幅缩小等离子体体积,单级能量增益随模式数线性增长,为紧凑型高能电子加速器开辟新路径。
Easy Reads:让arXiv论文告别「蚂蚁字」,一键重排为舒适阅读版式
Easy Reads是一个开源Python工具,能自动从arXiv下载论文LaTeX源文件并重新编译——字号从默认10pt提升到12pt或14pt,支持双栏转单栏。不是PDF放大,而是从源头重构排版。对于每天在屏幕上啃密集论文的科研人员来说,这是个实用的阅读辅助工具。
Easy Reads:让arXiv论文不再难读的Python开源工具
Easy Reads是一款开源Python程序,能够自动从arXiv获取论文TeX源文件,通过调整字号和栏数等排版参数生成更易读的PDF版本。它解决了arXiv论文紧凑排版在屏幕上阅读困难的痛点,无需改变论文内容,只需一个arXiv URL即可一键生成大字版论文,显著提升阅读舒适度和可访问性。
Zero-VC:零前瞻流式语音转换——用说话人匿名化技术实现真正实时的声线变换
Zero-VC提出了一种全新的零前瞻流式语音转换方法,利用说话人匿名化(SA)作为扰动机制,解决了现有方法在音色泄露与音质保持之间的权衡难题。该方法采用严格因果架构,完全无需缓冲未来帧,实现了真正的零延迟实时语音转换,被Interspeech 2026接收。
Zero-VC:零前瞻延迟的流式语音转换——用说话人匿名化突破实时变声的瓶颈
Zero-VC提出了一种零前瞻延迟的流式语音转换方法,创新性地将说话人匿名化(SA)作为扰动机制,在严格因果架构下实现高质量实时语音转换。该方法通过SA产生的鲁棒表征消除了模型对未来帧的依赖,在保持韵律完整性的同时避免音色泄漏,已被Interspeech 2026接收。
几何不一致多能谱CT中的高效精确图像重建:突破射线依赖能谱的重建瓶颈
研究团队提出了一种针对几何不一致、射线依赖能谱的多能谱CT图像重建新方法。通过引入聚合能谱概念并利用雅可比矩阵的特殊块结构(对角矩阵×极小规模矩阵),该算法在精度和效率上远超现有方法,为多能谱CT临床应用扫清了关键技术障碍。
量子生命游戏建模意大利西尼罗河病毒爆发:当元胞自动机遇上传染病学
意大利研究团队创造性地将量子版本的生命游戏元胞自动机应用于传染病建模,成功模拟了2025年意大利西尼罗河病毒爆发。该模型仅需优化蚊子出生率和移除率两个参数,即可高精度拟合实际感染曲线,并可量化评估灭蚊措施和气候变化对疫情的影响。
多智能体与多体系统的最优秩序:当个体力量遇上集体智慧
本文深入解读Jake J. Xia的最新研究,该框架通过力量和响应函数两个核心变量分析多智能体系统,揭示了总功率、有用功率、熵、秩序、脆弱性、流动性等宏观性质如何从异质个体中涌现,并引入风险偏好系数推导出平衡生产力、稳定性和适应性的最优秩序度。核心发现:同步性是双刃剑,秩序、熵和信息都是任务依赖的概念。
大规模随机基因调控网络中的振荡与空间图案:噪声如何成为生命的图案设计师
研究揭示分子噪声在基因调控网络图案形成中的关键作用。通过发展二阶矩随机分析框架,发现小系统中噪声可诱导图灵不稳定性,且无需不同物质扩散速率差异——这从根本上挑战了经典图灵理论的核心假设,为理解胚胎发育等生物过程中的空间图案形成提供了全新视角。
用神经网络补全大脑:混合生物物理神经元模型如何用Neural ODEs发现未知离子通道动力学
德国蒂宾根大学团队提出一种混合建模框架,将神经常微分方程嵌入Hodgkin-Huxley型生物物理模型,从电压记录中学习未知离子通道动力学和树突轴向电流。该方法在2400种离子通道模型上验证成功,能从单次电流钳记录中恢复未知门控动力学,还将多区室皮层神经元模型压缩为单区室模型,计算速度提升一个数量级。
量子计算突破:全活化空间波函数的高效经典表示与量子态制备
量子计算在化学领域的重大突破:研究者利用量子Paldus变换证明,全活化空间(CAS)波函数可以高效地表示为矩阵乘积态,键维数仅O(d²)增长。基于此设计的量子态制备算法将门复杂度从指数级降至O(d³)多项式级,实现了指数级改进。该成果对量子化学、药物设计和材料科学具有深远意义。
深度Transformer中的自适应推理与功能向量:从均场理论理解大模型如何学习
本文深入解读Ravin Raj和Gautam Reddy的最新研究,该工作从均场相互作用系统的角度重新审视深度Transformer,揭示了模型如何通过功能向量(function vectors)实现分布式推理,在层与层之间逐步推断潜在上下文变量。研究发现前馈模块和网络深度使Transformer能实现远比此前描述更丰富的上下文学习算法。
量子计算突破:完全活性空间波函数的高效经典表示与量子态制备
量子计算机有望解决复杂分子的电子结构问题,但强关联分子的量子态制备一直是瓶颈。本文解读Hamza Jnane的最新研究:利用量子Paldus变换(QPT),将完全活性空间(CAS)态在对称性适配基下展开为矩阵乘积态(MPS),使经典表示和量子态制备的复杂度从指数级降至多项式级O(d³),实现了对现有方法的指数级改进。
BrainWorld:用结构MRI引导全脑4D fMRI动态生成的全新框架
BrainWorld是一种结构先验条件生成模型,利用sMRI作为解剖学上下文来引导全脑4D fMRI动态生成。该模型在22个数据集上验证,可生成长达400帧的稳定fMRI轨迹,并通过生成样本增强提升下游任务性能,为脑动态建模和多模态表征学习开辟新路径。
揭秘婴儿大脑发育:结构-功能模块一致性如何从0到5岁演变
引入随机模块概念,提出鲁棒的概率性方法测量婴儿大脑结构-功能模块一致性,发现0-5岁期间SFMC持续下降,初级脑区一致性高于高级认知区域。
2026 DevTool Landscape: 6 New Tools Deep Dive
Deep dive into Cursor CLI, Warp Terminal, Cline, Kimi K2.6, Visual Studio 2026, and Aider - 6 latest developer tools of 2026 with features, tech stacks, use cases, and installation guides.
神经元会说话吗?单细胞分辨率下的视觉语义叙述
NEURRATOR框架首次实现了从单个神经元的放电活动中解码出自然语言描述的视觉场景,为理解大脑视觉编码提供了革命性工具。
Oracle PeopleSoft零日漏洞CVE-2026-35273深度解析:ShinyHunters与Cl0p联手发动大规模数据窃取攻击
CVSS 9.8分Oracle PeopleSoft零日漏洞CVE-2026-35273已被ShinyHunters和Cl0p两大勒索组织在野利用,全球高校和企业面临严重数据泄露风险。SSRF到RCE攻击链完整分析与修复方案。
2026 DevTool Roundup: 6 AI Coding Tools
Top 6 developer tools in 2026: Gemini CLI, OpenCode, Cline, VS 2026, Aider, Cursor
CVE-2026-10520: Ivanti Sentry OS Command Injection Root RCE
CVE-2026-10520 (CVSS 9.8) is an unauthenticated OS command injection in Ivanti Sentry allowing root-level RCE. CISA KEV added June 11. CrowdSec confirms active exploitation. Affects versions before R10.5.2, R10.6.2, R10.7.1.
GlassWorm Supply Chain Attack: 433 Components Compromised with Invisible Unicode
GlassWorm campaign (March 2026) infected 433 components across GitHub, npm, and VSCode/OpenVSX using invisible Unicode characters to hide malicious code. Stole Git/npm/GitHub credentials and drained crypto wallets. Largest coordinated cross-platform supply chain attack to date.
Nightmare-Eclipse: Windows Defender Six Zero-Day Chain Deep Analysis
Nightmare-Eclipse disclosed 6 Windows Defender zero-days (BlueHammer CVE-2026-33825, RedSun, UnDefend, YellowKey, GreenPlasma) in April-June 2026. Only BlueHammer patched. Ransomware gangs actively chaining these exploits to disable Defender and deploy payloads.
CVE-2026-20253: Splunk Enterprise Pre-Auth RCE via PostgreSQL Sidecar
CVE-2026-20253 (CVSS 9.8) allows unauthenticated attackers to achieve RCE on Splunk Enterprise through the PostgreSQL Sidecar service endpoint. CISA added to KEV on June 18, 2026. Only 5 HTTP requests needed for full compromise. Patch immediately.
CVE-2026-32746:GNU telnetd远程代码执行漏洞深度分析与修复指南
2026年3月披露的CVE-2026-32746是GNU InetUtils telnetd中隐藏近30年的关键缓冲区溢出漏洞,影响所有≤2.7版本。未经认证的远程攻击者可通过LINEMODE SLC处理器实现任意代码执行,全球150万台暴露telnet端口的设备面临风险。
CVE-2026-26956 vm2 Sandbox Escape via WebAssembly
CVSS 9.8 vm2 sandbox escape via WebAssembly JSTag exception handling, affecting 1.3M weekly npm downloads.
CVE-2026-22778-vLLM-RCE-via-Malicious-Video
OX Security discovers CVSS 9.8 RCE in vLLM via malicious video deserialization, affecting thousands of production LLM deployments.
Axios-npm-Supply-Chain-Attack-2026-Deep-Analysis
Axios npm package with 100M weekly downloads compromised via maintainer account hijack, RAT backdoor planted by North Korean APT UNC1069.
CVE-2026-3854:GitHub跨租户远程代码执行漏洞深度解析——一条git push拿下整个平台
GitHub.com和GitHub Enterprise Server曝出CVSS 9.8严重漏洞CVE-2026-3854,认证用户可通过单次git push获取远程代码执行权限,影响数百万代码仓库。本文深入分析漏洞原理、攻击链、检测方法和修复方案。
CVE-2026-48303:Adobe Campaign Classic CVSS满分10.0漏洞深度解析——营销自动化平台的致命后门
NVD披露Adobe Campaign Classic的CVSS 10.0满分漏洞CVE-2026-48303。该授权绕过漏洞允许远程攻击者以当前用户身份执行任意代码。ACC被财富500强中200+企业用于客户数据管理和多渠道营销,一旦被攻陷将导致大规模客户数据泄露。
Windows内核组件三连击:TCP/IP堆溢出、DHCP栈溢出、HTTP.sys整数溢出深度解析
2026年6月微软披露三个影响Windows核心网络组件的严重漏洞:CVE-2026-42904(TCP/IP堆溢出CVSS 9.6)、CVE-2026-44815(DHCP栈溢出CVSS 9.8)、CVE-2026-47291(HTTP.sys整数溢出CVSS 9.8)。三个漏洞均允许未经认证的远程代码执行,影响Windows最基础的网络协议栈。
CVE-2026-10520:Ivanti Sentry命令注入漏洞深度解析——从远程请求到完全控制
CISA将Ivanti Sentry命令注入漏洞CVE-2026-10520纳入KEV目录。该漏洞允许未经身份验证的远程攻击者通过构造特制HTTP请求执行任意操作系统命令。这是Ivanti产品在2026年被纳入CISA KEV的第3个漏洞,Ivanti已成为攻击者最青睐的目标之一。
CVE-2026-5281:Chrome WebGPU零日漏洞深度解析——Dawn组件UAF已在野利用
Google紧急修复2026年第4个Chrome零日漏洞CVE-2026-5281。Dawn组件UAF已在野利用,攻击者可通过恶意网页实现任意代码执行。
CVE-2026-8421:Concrete CMS CSRF到RCE漏洞深度解析
CVE-2026-8421是Concrete CMS(≤9.5.0)中发现的CSRF漏洞,攻击者可通过诱导已认证管理员访问恶意页面实现远程代码执行。本文深入分析攻击链、检测方法与修复方案。
CVE-2026-23010:Linux内核IPv6 use-after-free漏洞深度解析与修复指南
syzbot在Linux内核inet6_addr_del()函数中发现严重use-after-free漏洞(CVE-2026-23010),可导致内核崩溃甚至远程代码执行,影响所有启用IPv6的Linux服务器。本文从漏洞原理、攻击面分析、检测方法到修复方案进行全方位解读。
CVE-2026-26956:vm2沙箱逃逸漏洞深度解析——WebAssembly绕过Node.js隔离(CVSS 9.8)
CVE-2026-26956(CVSS 9.8)是Node.js沙箱库vm2的严重逃逸漏洞,攻击者通过WebAssembly绕过沙箱隔离实现RCE,每周130万次下载的npm包受影响。本文详解攻击原理、PoC代码及防御方案。
CVE-2026-7482 Ollama Bleeding Llama:CVSS 9.1的AI基础设施内存泄露漏洞深度解析(2026)
Ollama被曝CVSS 9.1严重漏洞CVE-2026-7482,未认证攻击者可泄露服务器全部进程内存,全球超30万台服务器受影响。本文深度剖析漏洞原理、PoC利用链、影响范围及完整修复方案。
CVE-2026-21858:n8n「Ni8mare」漏洞——CVSS满分10.0的工作流自动化平台RCE
CVE-2026-21858(CVSS 10.0)允许无认证攻击者通过Content-Type混淆实现n8n服务器完全控制。PoC已公开,全球约10万台实例面临风险。
2026年6月最值得关注的6款开发者工具:从AI编码到项目管理的全面升级
深度解析Gemini CLI、Kilo、Cline、VS 2026、moonrepo、Plane六款开发者工具,涵盖AI编码代理、构建系统、项目管理,附代码示例与对比表格
apple/coreai-models:苹果开源AI模型,推动端侧智能革命
apple/coreai-models是苹果开源的AI模型导出和运行时工具,支持iPhone/iPad/Mac全设备部署。本文深度解析其技术架构、快速上手指南、应用场景、变现方式和学习路径,助你掌握Apple端侧AI开发。
自复制AI蠕虫攻击全面解析:开源大模型新型威胁与防御策略(2026)
研究人员构建出首个完全在本地开源大模型上运行的自复制AI蠕虫。深度解析AI蠕虫传播机制、攻击原理、检测方法及AI系统安全防御策略。
Veeam Backup远程代码执行漏洞深度分析:域用户RCE威胁与防御实战(2026)
Veeam Backup & Replication曝远程代码执行漏洞,域用户可直接执行恶意代码。详解漏洞原理、影响范围、检测方法及企业备份安全加固方案。
LiteLLM严重漏洞CVE-2026-42271:AI代理框架未授权RCE全解析(2026)
AI代理框架LiteLLM曝严重漏洞CVE-2026-42271,攻击链可实现未授权远程代码执行。深度分析漏洞原理、利用方式、修复方案及AI安全防御策略。
Dedalus Labs深度解析:YC S25最新项目,做AI Agent的Vercel,创业机会在哪里?
# Dedalus Labs深度解析:YC S25最新项目,做AI Agent的Vercel,创业机会在哪里?  > Y Com
Anima:让每个硬件都有灵魂,Fullive AI开源的智能家居Agent OS深度解析
# Anima:让每个硬件都有灵魂,Fullive AI开源的智能家居Agent OS深度解析  > 当你的空调不再只是听命行事
JoyAI-Echo深度解析:京东开源的5分钟长视频生成AI,突破音视频生成新边界(2026)
# JoyAI-Echo深度解析:京东开源的5分钟长视频生成AI,突破音视频生成新边界(2026)  > 2026年6月,京东开
TripoSplat完全解析:一张照片秒变3D模型,VAST AI开源的高斯溅射神器(2026)
# TripoSplat完全解析:一张照片秒变3D模型,VAST AI开源的高斯溅射神器(2026)  > 2026年6月,VA
2026年最值得学习的5个开源项目
从AI框架到数据库,这些开源项目正在改变技术格局
开源大模型本地部署实战:Llama 4 vs Qwen 3 vs DeepSeek V3
三大开源模型本地部署性能横评,附硬件需求和实测数据
2026年AI编程助手终极对决:GitHub Copilot vs Cursor vs Cline
三大AI编程助手横评:代码补全、多文件编辑、Agent模式全面对比
OpenHack深度解析:Hadrian Security开源的白盒安全审查工具完全指南
OpenHack是由Hadrian Security开源的源码导向白盒安全审查工具。本文深度解析其架构、使用方法、与传统SAST工具的对比,以及在实际项目中的应用实践。
AIDA:将任何LLM变成自主渗透测试代理的开源安全工具
AIDA是开源自主渗透测试代理,可将Claude、GPT等LLM变成自动化安全评估工具。已发现真实CVE漏洞,Docker隔离执行。
HyperFrames深度解析:HeyGen开源的HTML转视频框架,AI代理时代的视频生成利器
HeyGen开源的HyperFrames将HTML/CSS/动画渲染为MP4视频,专为AI代理设计。25K+ Star,支持CLI、AI技能和可视化设计。本文深度解析技术架构、六大应用场景与变现路径。
GWS CLI:Google官方开源的Google Workspace命令行工具
Google Workspace官方CLI工具gws,一条命令管理Drive、Gmail、Calendar等所有Workspace服务。Rust实现,动态API发现,内置40+ AI代理技能。
Agent Browser:Vercel开源的AI代理浏览器自动化CLI工具
Vercel Labs开源的Agent Browser是专为AI代理设计的浏览器自动化CLI工具,基于Rust构建,速度快、功能强。35K+ Star。
Nanobot:香港大学开源的超轻量级个人AI代理,重新定义智能助手
HKUDS团队开源的Nanobot是超轻量级个人AI代理,支持工具调用、多轮对话和工作流自动化。43K+ Star,Python实现,MIT协议。
Karpathy Autoresearch深度解析:让AI代理自主进行深度学习研究的革命性开源工具
Karpathy开源的Autoresearch让AI代理在单GPU上自主运行深度学习实验,自动修改代码、训练模型、评估结果。85K+ Star,开启AI自主研究新时代。本文深度解析三文件架构、自主研究循环与对AI研究范式的影响。
2026年开发者必看:NVIDIA NemoClaw开源、AI调试新工具、28款顶级安全工具盘点
NVIDIA开源企业级AI Agent平台NemoClaw、Raindrop发布本地AI Agent调试工具Workshop、Wiz盘点2026年28款最佳开源安全工具——开发者工具生态持续爆发。
AI行业重磅:特朗普签署AI行政令、NVIDIA开源物理AI工具、Gemma 4发布
特朗普签署AI行政令要求企业向政府开放模型访问、NVIDIA发布Physical AI开源工具集、Google推出Gemma 4 12B多模态模型——AI行业政策与技术双线并进。
Anima: 让每个硬件都有灵魂,Fullive AI开源的智能家居Agent OS
当你的空调不再只是一个听命行事的执行器,而是一个能理解"有点热"、记住你习惯、主动为你调节环境的智能体——这就是Anima想要实现的未来。
Sandboxed: 开源AI应用构建引擎一行命令搭建Lovable级别开发环境
Sandboxed是开源AI应用构建后端引擎,一行命令创建隔离云开发环境。内置AI编程代理,支持实时预览URL,无需Kubernetes。Go语言实现,智能休眠唤醒。
TripoSplat: 一张照片秒变3D模型VAST AI开源高斯溅射神器
TripoSplat是VAST AI开源的单图转3D工具,将2D照片实时转换为高质量3D高斯溅射模型。支持可变数量高斯基元,适用于游戏资产、AR/VR和电商场景。
JoyAI-Echo深度解析:京东开源5分钟长视频生成AI,突破音视频生成新边界
京东开源JoyAI-Echo视频生成模型,支持5分钟连贯长视频+音视频一体化生成。本文深度解析技术架构、与Sora/Kling对比、部署方案与变现路径。
Skylight: 把头顶飞过的飞机投射到天花板上的开源神器
Skylight是一个革命性的开源实时航空可视化系统,通过RTL-SDR接收ADS-B信号,将头顶实时飞过的飞机以沉浸式投影方式呈现在天花板上。支持60fps平滑动画、类型感知飞机图标、真实星空背景,打造属于你的私人'天窗'体验。
CVE-2026-47117漏洞深度解析:OpenMed远程代码执行漏洞技术分析与防御策略
CVE-2026-47117是OpenMed医疗信息系统中的远程代码执行漏洞(CVSS 9.8),通过JNDI注入实现未授权RCE。本文从攻击者视角剖析漏洞成因、利用链构造和企业级防御方案。
Sandboxed深度解析:AI应用构建的开源沙箱引擎(2026最新)
全面解析Sandboxed沙箱引擎的技术架构、Docker隔离机制和预览URL系统。为AI编码代理提供安全的运行环境,一条命令部署,无需Kubernetes。
Skylight深度解析:将飞机投影到天花板的开源神器完全指南
Skylight是一个将实时飞机航班投影到天花板的开源项目,让你在家就能看到头顶飞过的飞机。本文深度解析其原理、安装、配置和使用体验。
CVE-2026-45247:Magento RCE漏洞已被CISA列入KEV
CISA将Magento扩展CVE-2026-45247远程代码执行漏洞加入已知被利用漏洞目录影响Mirasvit Cache Warmer插件。详解漏洞详情、野外利用方式和紧急修复方案。
OpenLogi:Rust打造的Logitech Options+开源替代方案
OpenLogi是用Rust编写的Logitech Options+开源替代品,支持按钮重映射、DPI调节、SmartShift,无需账户无遥测,本地TOML配置,跨平台Windows/macOS/Linux。
Odysseus自托管AI工作空间完整指南(2026最新版)
全面解析Odysseus自托管AI工作空间,支持vLLM、Ollama、OpenAI多后端,集成聊天、Agent、深度研究、记忆、邮件、日历全功能,Docker一键部署,本地优先隐私保护的AI解决方案。
云微信WechatOnCloud完全指南:1.9K星的微信云化方案深度评测
WechatOnCloud是一个将微信运行在云端的开源方案,支持远程访问和多设备同步。本文深度解析其架构、部署、安全性和实际使用体验。
Odysseus完全解析:41K星的自托管AI工作空间深度指南
Odysseus是一个开源的自托管AI工作空间,支持多AI模型、插件系统和团队协作。本文深度解析其架构、部署、使用和最佳实践。
🐙 今日开源精选 | 2026-05-29
SQLite持久化工作流、Zig构建系统重构、TypeScript直接编译为可执行文件、PrusaSlicer多色打印开源方案——今日开源社区四大热门项目。
情报动态 200
elementalsouls/Claude-BugHunter(16,100★)
Claude 代码技能包,用于漏洞狩猎和外部红队工作,包含 82 个技能、15 个斜杠命令和 681 个公开报告模式,对安全开发者有价值
google-labs-code/stitch-skills(11,450★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
mvanhorn/last30days-skill(72,015★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
firecrawl/firecrawl(206,223★)
Knowledge infrastructure to make your agents smarter. Search, scrape and access 550+ data providers through one API.
jingyaogong/minimind(68,160★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
HKUDS/CLI-Anything(62,085★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
obra/superpowers(323,738★)
An agentic skills framework & software development methodology that works.
elementalsouls/Claude-BugHunter(16,092★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
mjmirza/app-store-compliance(3,308★)
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.
google-labs-code/stitch-skills(11,366★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
mvanhorn/last30days-skill(71,911★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
cathrynlavery/diagram-design(69,748★)
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, Qoder, and Pi. 45 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.
openvinotoolkit/openvino(11,908★)
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
firecrawl/firecrawl(206,068★)
Knowledge infrastructure to make your agents smarter. Search, scrape and access 550+ data providers through one API.
jingyaogong/minimind(68,098★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
HKUDS/CLI-Anything(61,988★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
obra/superpowers(323,626★)
An agentic skills framework & software development methodology that works.
elementalsouls/Claude-BugHunter(15,810★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
mjmirza/app-store-compliance(3,279★)
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.
google-labs-code/stitch-skills(11,349★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
mvanhorn/last30days-skill(71,892★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
cathrynlavery/diagram-design(69,729★)
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, Qoder, and Pi. 45 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.
openvinotoolkit/openvino(11,900★)
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
firecrawl/firecrawl(206,017★)
Knowledge infrastructure to make your agents smarter. Search, scrape and access 550+ data providers through one API.
jingyaogong/minimind(68,083★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
HKUDS/CLI-Anything(61,957★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
obra/superpowers(323,597★)
An agentic skills framework & software development methodology that works.
elementalsouls/Claude-BugHunter(15,735★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
mjmirza/app-store-compliance(2,646★)
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.
google-labs-code/stitch-skills(10,981★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
mvanhorn/last30days-skill(71,374★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
cathrynlavery/diagram-design(69,219★)
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, Qoder, and Pi. 45 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.
openvinotoolkit/openvino(11,756★)
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
firecrawl/firecrawl(205,140★)
Knowledge infrastructure to make your agents smarter. Search, scrape and access 550+ data providers through one API.
jingyaogong/minimind(67,643★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
HKUDS/CLI-Anything(61,295★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
obra/superpowers(322,770★)
An agentic skills framework & software development methodology that works.
elementalsouls/Claude-BugHunter(13,998★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
mjmirza/app-store-compliance(1,798★)
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.
google-labs-code/stitch-skills(10,699★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
mvanhorn/last30days-skill(70,768★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
cathrynlavery/diagram-design(68,605★)
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, Qoder, and Pi. 45 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.
openvinotoolkit/openvino(11,549★)
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
firecrawl/firecrawl(204,156★)
Knowledge infrastructure to make your agents smarter. Search, scrape and 550+ data providers through one API.
jingyaogong/minimind(66,830★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
HKUDS/CLI-Anything(60,522★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
obra/superpowers(322,043★)
An agentic skills framework & software development methodology that works.
elementalsouls/Claude-BugHunter(12,452★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
mjmirza/app-store-compliance(1,749★)
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.|点评:App Store 与 Play 商店合规手册,防止拒绝
google-labs-code/stitch-skills(10,667★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.|点评:Stitch MCP 兼容的 Agent 技能库,支持多代理
mvanhorn/last30days-skill(70,716★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary|点评:AI 技能聚合多平台内容,生成扎实总结
cathrynlavery/diagram-design(68,558★)
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, Qoder, and Pi. 45 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.|点评:Claude 代码、Copilot 等工具的自包含 SVG 图表
openvinotoolkit/openvino(11,532★)
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference|点评:OpenVINO 工具包,优化 AI 推理部署
nginx/nginx(32,295★)
The official NGINX Open Source repository.|点评:NGINX 官方开源项目,成熟高性能 Web 服务器
firecrawl/firecrawl(204,047★)
Knowledge infrastructure to make your agents smarter. Search, scrape and 550+ data providers through one API.|点评:一键搜索抓取 550+ 数据源,构建知识基础设施
cursor/plugins(17,517★)
Cursor plugin specification and official plugins|点评:Cursor 插件规范与官方插件,扩展代理功能
jingyaogong/minimind(66,796★)
🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.|点评:两小时训练 64M LLM,快速构建小型语言模型
HKUDS/CLI-Anything(60,467★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/|点评:让所有软件具备原生 CLI,统一命令行体验
obra/superpowers(321,985★)
An agentic skills framework & software development methodology that works.|点评:面向代理的技能框架与开发方法,提升软件开发效率
elementalsouls/Claude-BugHunter(12,336★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.|点评:Claude 代码技能包,覆盖 24 类漏洞,适合红队与漏洞挖掘
google-labs-code/stitch-skills(10,371★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.|点评:Agent 技能库,兼容多种编码代理,助力自动化开发
google-research/paper-orchestra(4,166★)
PaperOrchestra is a multi-agent framework for automated AI research paper writing.|点评:多智能体框架,自动化 AI 论文写作,科研效率翻倍
astral-sh/uv(92,199★)
An extremely fast Python package and project manager, written in Rust.|点评:Rust 编写的 Python 包管理器,速度快、占用低,提升 Python 开发体验
nvm-sh/nvm(96,145★)
Node Version Manager - POSIX-compliant bash script to manage multiple active node.js versions.|点评:Node 版本管理器,POSIX 脚本,轻松切换多版本,开发环境必备
NVIDIA/SkillEvaluator(3,656★)
Multi-tier framework for evaluating AI agent skills and plugins with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.|点评:多层 AI 评估框架,检测技能质量与重叠,支持实时评测
getarcaneapp/arcane(9,762★)
Modern Docker Management, Designed for Everyone|点评:现代 Docker 管理工具,简化容器部署,面向所有开发者
mvanhorn/last30days-skill(69,898★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary|点评:AI 代理技能,实时检索社交媒体与网络,生成可靠摘要,适合信息采集
HKUDS/CLI-Anything(59,623★)
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/|点评:将所有软件转为 CLI,CLI-Hub 提供统一入口,极大提升自动化效率
cursor/plugins(17,157★)
Cursor plugin specification and official plugins|点评:Cursor 插件规范与官方插件,统一标准,方便开发者快速集成
elementalsouls/Claude-BugHunter(11,175★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.|点评:Claude 代码技能包,覆盖 24 类漏洞,适合红队与漏洞挖掘,值得关注
backnotprop/plannotator(13,015★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, and send all feedback to agents with one click.
ag-ui-protocol/ag-ui(19,925★)
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.
obra/superpowers(319,224★)
An agentic skills framework & software development methodology that works.
google-labs-code/stitch-skills(10,045★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
dream-num/univer(28,840★)
Build your own AI office with Univer Office SDK. Bring editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.
Tencent/LoopForge(3,040★)
Self-hosted workflows for coding-agent delivery. Orchestrate CodeBuddy, Codex and Claude Code with scripts and human approval gates. Includes Local workflows for coding assistants.
alibaba/open-code-review(53,647★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
mvanhorn/last30days-skill(68,911★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
elementalsouls/Claude-BugHunter(8,994★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
Building the Agent Host Protocol for sessions that continue across devices
Agent Host Protocol实现会话跨设备无缝衔接
alibaba/open-code-review(52,872★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
NVIDIA/SkillEvaluator(2,550★)
Multi-tier framework for evaluating AI agent skills and plugins with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.
mvanhorn/last30days-skill(68,043★)
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
google-labs-code/stitch-skills(9,689★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
DeanChensj/jiangshan-map(6,653★)
江山时序 · 中国历史交互地图时间轴 (403 BCE – Present) | Interactive Historical Atlas of China across 18 dynasties, 198 territory snapshots, ancient-to-modern place genealogy & classical poetry routes
elementalsouls/Claude-BugHunter(7,299★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
tradermonty/claude-trading-skills(6,221★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners, and trading strategy development.
dream-num/univer(27,707★)
Build your own AI office with Univer Office SDK. Bring editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.
backnotprop/plannotator(12,464★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, and send all feedback to agents with one click.
ag-ui-protocol/ag-ui(18,833★)
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.
NVIDIA/SkillEvaluator(1,576★)
Multi-tier framework for evaluating AI agent skills with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.
tradermonty/claude-trading-skills(5,431★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners, and trading strategy development.
GoogleChrome/modern-web-guidance-src(7,028★)
Keep your coding agent up to date with the latest web best practices
elementalsouls/Claude-BugHunter(6,284★)
A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
Tencent/LoopForge(1,399★)
Self-hosted workflows for coding-agent delivery. Orchestrate CodeBuddy, Codex and Claude Code with scripts and human approval gates. Includes Local workflows for coding assistants.
alibaba/open-code-review(51,948★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
backnotprop/plannotator(11,761★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, and send all feedback to agents with one click.
google-labs-code/stitch-skills(9,232★)
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
ag-ui-protocol/ag-ui(17,772★)
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.
dream-num/univer(25,518★)
Build your own AI office with Univer Office SDK. Bring editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.
JuliusBrussee/caveman(119,997★)
🪨 why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.|点评:Caveman代理,低token化沟通,显著节省成本。
ashhart/TensorFold(2,144★)
LLM Inference Engine for Metal, CUDA and Vulkan.|点评:TensorFold推理引擎,兼容Metal/CUDA/Vulkan,性能卓越。
ag-ui-protocol/ag-ui(17,309★)
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.|点评:AG-UI协议,简化Agent与前端交互,提升体验。
dream-num/univer(24,418★)
Build your own AI office with Univer Office SDK. Bring editing, calculation, collaboration, and file processing into your product for people and AI agents. Built on an extensible open-source core.|点评:Univer SDK打造AI办公套件,集成编辑与协作。
ChromeDevTools/chrome-devtools-mcp(56,224★)
Chrome DevTools for coding agents|点评:Chrome DevTools插件,支持编码代理高效调试。
tradermonty/claude-trading-skills(4,717★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners, and trading strategy development.|点评:Claude交易工具,提供市场分析与策略开发。
GoogleChrome/modern-web-guidance-src(5,864★)
Keep your coding agent up to date with the latest web best practices|点评:现代Web最佳实践源码,保持前端代码更新。
backnotprop/plannotator(11,020★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, and send all feedback to agents with one click.|点评:Agent协作平台,集成注释、评审与分享,提升团队效率。
tradermonty/claude-trading-skills(4309★,今日+69★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners
tailscale/tailcat(9575★,今日+256★)
like netcat, but over Tailscale's data plane, without Tailscale's control plane
backnotprop/plannotator(10565★,今日+108★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, an
ZhuLinsen/daily_stock_analysis(71873★,今日+256★)
LLM 驱动的多市场股票智能分析系统:多源行情、实时新闻、决策看板与自动推送,支持零成本定时运行。 LLM-powered multi-market stock analysis system with multi-source marke
GoogleChrome/modern-web-guidance-src(5085★,今日+1925★)
Keep your coding agent up to date with the latest web best practices
openai/codex-security(13270★,今日+751★)
OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https:
alibaba/open-code-review(50847★,今日+215★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines
tailscale/tailcat(9498★,今日+256★)
like netcat, but over Tailscale's data plane, without Tailscale's control plane
tradermonty/claude-trading-skills(4217★,今日+69★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners
backnotprop/plannotator(10501★,今日+108★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, an
ZhuLinsen/daily_stock_analysis(71694★,今日+256★)
LLM 驱动的多市场股票智能分析系统:多源行情、实时新闻、决策看板与自动推送,支持零成本定时运行。 LLM-powered multi-market stock analysis system with multi-source marke
GoogleChrome/modern-web-guidance-src(4945★,今日+1925★)
Keep your coding agent up to date with the latest web best practices
openai/codex-security(13176★,今日+751★)
OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https:
alibaba/open-code-review(50710★,今日+215★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines
tailscale/tailcat(9487★,今日+256★)
like netcat, but over Tailscale's data plane, without Tailscale's control plane
tradermonty/claude-trading-skills(4205★,今日+69★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners
backnotprop/plannotator(10490★,今日+108★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, an
ZhuLinsen/daily_stock_analysis(71669★,今日+256★)
LLM 驱动的多市场股票智能分析系统:多源行情、实时新闻、决策看板与自动推送,支持零成本定时运行。 LLM-powered multi-market stock analysis system with multi-source marke
openai/codex-security(13166★,今日+751★)
OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https:
alibaba/open-code-review(50675★,今日+215★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines
GoogleChrome/modern-web-guidance-src(4922★,今日+1925★)
Keep your coding agent up to date with the latest web best practices
Likely7/Veyra-NRVideo(1,574★)
Bring PS5 GTA6 to 120FPS with DLSS 5. An advanced Windows video player and capture enhancement tool featuring real-time preview for videos, images, and capture card sources. Combine AI-powered Super Resolution, DLSS5 image enhancement, and Frame Generation to deliver a next-gener
tradermonty/claude-trading-skills(4,098★)
Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners, and trading strategy development.
backnotprop/plannotator(10,393★)
The workspace to stay in the loop and on top of agents - shape, decide, plan, annotate, review code, share with team, and send all feedback to agents with one click.
ZhuLinsen/daily_stock_analysis(71,510★)
LLM 驱动的多市场股票智能分析系统:多源行情、实时新闻、决策看板与自动推送,支持零成本定时运行。 LLM-powered multi-market stock analysis system with multi-source market data, real-time news, decision dashboard, automated notifications, and cost-free scheduled runs.
openai/codex-security(13,057★)
OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security
GoogleChrome/modern-web-guidance-src(4,676★)
Keep your coding agent up to date with the latest web best practices
alibaba/open-code-review(50,513★)
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
Show HN: Isometric SF – A pixel-art map of the SF Bay Area
Technical Write-up: https://www.justinchavez.com/blog/isometric-sf-how/ Hi HN! Showcasing this fun project. I took 1300 square miles of the SF Bay Area from Google Photorealistic 3D Tiles and transformed them into a cohesive pixel map. It's really fun to explore and stumble upon various famous landm
GitHub availability report: October 2026
In October, we experienced eight incidents that resulted in degraded performance across GitHub services. The post GitHub availability report: October 2026 appeared first on The GitHub Blog.
Alishahryar1/free-claude-code(69,179★)
Use Claude Code, Codex, Antigravity, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS frie
CopilotKit/OpenBot(11,502★)
Open-source AI coworkers that each get a computer of their own: a browser, files and tools, with every action decided before it happens and recorded after. Bring any AG-UI agent.
meta/astryx(26,380★)
An open source design system that's fully customizable and agent ready
addyosmani/agent-skills(116,596★)
Production-grade engineering skills for AI coding agents.