AI资讯大全AIPPXP.CN搜索 ↗
Updated daily · 100% free每日自动更新 · 全部免费下载

Great AI tools,
right in your hands.
把好用的 AI,
直接放到你手里。

News, tutorials, extensions, tools and skill packs — six sections refreshed every day.
See what changes, and take home something you can use right now.

资讯 · 教程 · 插件 · 工具 · 技能包,六大板块每天上新。
看见变化,也拿到能立刻上手的东西。

6每日更新板块sections daily
231+精选栏目内容curated items
111+真实文件下载real downloads
100%免费无套路free, no tricks
学 · 用 · 探索Learn · Use · Exploreai+保持好奇,持续探索。Stay curious, keep exploring.
📰 AI 资讯📘 AI 教程🧩 AI 插件🛠 AI 工具🎁 技能包⬇ 下载专区
🌅今日一句 · Daily line:每天进步一点点 —— 每天进步一点点,每天读一点、练一点,AI 时代你就是领跑的那一个。One small step every day beats a big plan someday.去练一练 →
AI 快讯

AI 快讯AI News · Updated daily每日自动更新 · 2026年10月12日

查看全部 150 条 →
正能量公益广告今日正能量每天进步一点点,AI 陪你把想法变成现实。从一篇图文教程开始 →广告
AI 教程

AI 教程AI Tutorials · Step-by-step guides you can follow手把手图文教程,跟着做就会

查看全部 62 条 →
图文教程 · 手把手

A Step-by-Step Guide to Setting Up Claude BugHunter: Complete Your First Authorised Vulnerability Test from Scratch手把手教你装好 Claude BugHunter:从零跑通第一次授权漏洞测试

In our previous article, we introduced the Claude BugHunter skill pack, which has risen to the top of GitHub Trending today. In this article, we will take you through the entire process, from downloading and installing the pack to conducting real-world practice on a live range: Equip yourself with 83 vulnerability hunting skills and complete your first authorised penetration test on DVWA and OWASP Juice Shop.

上一篇我们介绍了今天冲上 GitHub Trending 第一的 Claude BugHunter 技能包,这篇带你从下载、安装到真实靶场实战完整走一遍:装好 83 个漏洞猎杀技能,在 DVWA 和 OWASP Juice Shop 上完成你的第一次授权渗透测试。

图文教程 · 手把手

A Step-by-Step Guide to Using CC Switch: Manage All AI Programming Tool Providers and Configurations in 10 Minutes手把手教你用 CC Switch:10 分钟管好所有 AI 编程工具的供应商与配置

In the previous article, we introduced the CC Switch desktop manager, which has received 156,000 stars. This article will guide you from downloading and installing the software to completing your first provider switch, and then to managing the MCP server and Skills in a unified manner. We will use the official portable version that was released today throughout the process. We will use Windows as an example, but the process is the same for macOS/Linux.

上一篇我们介绍了 15.6 万星的 CC Switch 桌面管理器,这篇带你从下载安装到完成第一次供应商切换,再把 MCP 服务器和 Skills 也统一管起来。全程用今天入库的官方便携版,Windows 为例,macOS/Linux 流程相同。

图文教程 · 手把手

A Step-by-Step Guide to Setting Up Agent Autoloop: Let Claude Code Handle Tasks in the Background手把手教你装好 Agent Autoloop:让 Claude Code 在后台自己接活

In the previous article, we introduced Agent Autoloop, the "Agent Automation Mid-End". In this article, we'll guide you through setting it up from scratch and going through the complete closed loop of "Assign Task → Run in the Background → Approve Card → Complete Task". The official source code package released today will be used throughout the process, and it is compatible with Windows, macOS, and Linux.

上一篇我们介绍了 Agent Autoloop 这台「Agent 自动化中台」,这篇带你从零装好它,并真实走一遍「派活 → 后台跑 → 卡片拍板 → 干完收工」的完整闭环。全程用今天入库的官方源码包,Windows、macOS、Linux 通用。

图文教程 · 手把手

A step-by-step guide to reviewing code locally with PR-Agent: No need to install a bot – generate a report with a single command手把手用 PR-Agent 本地审查代码:不装机器人,一条命令出报告

Teams may not be willing to install robots in their repositories, but everyone wants to merge good code. This tutorial uses the official PyPI package of PR-Agent v0.50.0, which was added to the repository today, to guide you through the four steps of "installing the package → configuring the key → reviewing a PR locally → understanding the report", without touching any platform configuration throughout the process.

团队未必愿意给仓库装机器人,但每个人都想把好代码关。这篇教程用今天入库的 PR-Agent v0.50.0 官方 PyPI 包,带你走通「装包 → 配 key → 本地审一个 PR → 读懂报告」四步,全程不碰任何平台配置。

AI 插件

AI 插件AI Extensions · Browser extensions with official packs浏览器插件,附官方安装包

查看全部 37 条 →
开源插件 · 免费用上前沿模型

dsh-our-free-model: Use DeepSeek V4.1 Flash and Kimi K3 for free without logging in to DeepSeek Harnessdsh-our-free-model:在 DeepSeek Harness 里免登录免费用 DeepSeek V4.1 Flash、Kimi K3

This is one of the AI projects that has gained the most stars in the last 7 days (No. 1 on the GitHub Trending daily list, with a total of 81,000 stars): Simply install a plugin in DeepSeek Harness (dsh), and you can directly use cutting-edge models such as DeepSeek V4.1 Flash and Kimi K3, without needing to log in, register, or enter any API key. It's completely free and unlimited. Pure plugin installation: No kernel modifications, no build steps, zero dependencies.

这是最近 7 天涨星最快的 AI 项目之一(GitHub Trending 日榜第一,累计 8.1 万星):只需往 DeepSeek Harness(dsh)里装一个插件,无需登录、注册或填任何 API Key,就能直接用上 DeepSeek V4.1 Flash、Kimi K3 等前沿模型,完全免费、不限量。纯插件挂载:不改内核、无构建步骤、零依赖。

开源 AI 安全 · 蓝方防御 MCP

BLA: An open-source Blue Team defence MCP plugin that equips software with "AI reverse engineering protection" armourBLA:给软件穿上「防 AI 逆向」护甲的开源蓝方防御 MCP 插件

An AI Reverse Engineering Agent (REA) can tear apart a release package in just a few minutes, while the defenders are still using obfuscators from the last century. BLA (Blind LLM Agents) is an open-source blue team defence MCP & CLI that rose to 5th place on the GitHub Trending daily list this week: 31 defence tools covering 26 product formats, using the three-pronged approach of "poisoning → exhaustion → hallucination" to bankrupt AI reverse engineering in terms of computing power.

AI 逆向 Agent(REA)几分钟就能把一个发布包拆个底朝天,防御方却还在用上世纪的混淆器。BLA(Blind LLM Agents)是本周冲上 GitHub Trending 日榜第 5 的开源蓝方防御 MCP & CLI:31 个防御工具、覆盖 26 种产物格式,用「投毒 → 耗尽 → 幻觉」三板斧让 AI 逆向在算力上破产。

谷歌官方 · 设计评审工作流

Elephant-Goldfish Model: Google's open-source "Elephant Drafting, Goldfish Reviewing" specification design plugin, 21.1K starsElephant-Goldfish Model:谷歌开源「大象起草、金鱼评审」规格设计插件,21.1K 星

The biggest pitfall when writing software specifications is "reviewing your own work": the context is too saturated, and blind spots cannot be seen. EGM uses the "Elephant" role to hold all the project context and draft the specifications, and then uses the "Goldfish" reviewer, who has zero context, to pick out flaws with fresh eyes and identify unstated assumptions and oversights. Supports mainstream coding agents such as Claude Code, Codex, and Antigravity.

写软件规格时最大的坑是「自己审自己」:上下文太满,盲区看不见。EGM 用「大象」角色握住全部项目上下文起草规格,再用零上下文的「金鱼」评审员带着新鲜眼光挑毛病,把没说出口的假设和遗漏揪出来。支持 Claude Code、Codex、Antigravity 等主流编码 Agent。

插件介绍 · 附完整插件包

Claude-Obsidian: Turn Claude Code and Obsidian into a "local-first" knowledge work partnership. Any information you put in becomes a note with a reference link.claude-obsidian:让 Claude Code 和 Obsidian 变成一对「本地优先」知识工作搭档,任何资料丢进去都变成带引用链接的笔记

Do you have two hundred unread articles piled up in your favourites and dozens of PDFs sitting in your downloads folder, and you want AI to help you process them but are always worried that it will make up sources? This open-source project with over 16,400 stars directly connects Claude Code (or any host compatible with Agent Skills) to the Obsidian vault: you feed it any source, and it generates interlinked Markdown notes with each conclusion citing its source. When you ask questions, the answers are strictly based on the evidence already in the vault, and if it can't find an answer, it simply says so. It runs entirely locally, and your data never leaves your computer.

收藏夹里堆了两百篇没读的文章、下载目录里躺着几十份 PDF,想让 AI 帮你消化却总担心它瞎编出处?这个 16,400+ 星的开源项目把 Claude Code(或任何兼容 Agent Skills 的宿主)和 Obsidian 库直接接通:你把任何来源丢进去,它产出互相链接、每个结论都标明出处的 Markdown 笔记;提问时回答也严格基于库里已有证据,不给答案就直说找不到。全程本地运行,数据不出你的电脑。

MIT 许可 · 16,400+ 星 · v3.0.1 · 本地优先查看 →
AI 工具

AI 工具AI Tools · Desktop & CLI tools, ready to download桌面与命令行工具,直接下载

查看全部 41 条 →
开发者工具 · 桌面端 · 15.6 万星

CC Switch: One-click switching of API providers for 10+ AI programming tools such as Claude Code and CodexCC Switch:一键切换 Claude Code、Codex 等 10+ AI 编程工具的 API 供应商

Who among those who use Claude Code, Codex, and Gemini CLI at the same time hasn't experienced the agony of spending an entire evening manually editing config.json? CC Switch (156,000 stars, built with Tauri 2) is a cross-platform desktop manager: Switch the API providers for all AI programming CLIs with one click, and centrally manage their MCP servers, Skills, and Prompts – no more manual editing of configuration files.

同时用 Claude Code、Codex、Gemini CLI 的人,谁没经历过手改一晚上 config.json 的痛苦?CC Switch(15.6 万星,Tauri 2 构建)是一个跨平台桌面管理器:一键切换所有 AI 编程 CLI 的 API 供应商,还能集中管理它们的 MCP 服务器、Skills 和 Prompts,不用再手动编辑任何配置文件。

国产开源 · Agent 自动化中台

Agent Autoloop: Let the command-line Agent handle tasks in the background on its own, and only come to you for decisions when it's unsure.Agent Autoloop:让命令行 Agent 后台自己接活,拿不准才来找你拍板

The biggest waste of a command-line agent is having someone man the terminal: if you walk away, it stops. Agent Autoloop is a Chinese open-source project that was only created on 3 November and reached 5,909 stars in just two weeks – an automation middleware with zero external dependencies. It turns event sources, agents, and notification channels into pluggable plugins. Tasks queue up and follow up on their own, and only when a decision is needed will a card be pushed to Lark, DingTalk, WeChat Work, or Slack to notify you.

命令行 Agent 最大的浪费是人守着终端:你走开,它就停。Agent Autoloop 是 11 月 3 日才创建、两周冲到 5909 星的国产开源项目——一个零外部依赖的自动化中台,把事件来源、Agent、通知通道全部做成可插拔插件,任务自己排队跟进,只有需要拍板时才把卡片推到飞书、钉钉、企微或 Slack 找你。

开源鼻祖 · AI 代码审查

PR-Agent: The pioneer of open-source AI code review is back, offering a health check for any PR with a single command.PR-Agent:开源 AI 代码审查的鼻祖回归,一行命令给任意 PR 做体检

AI code reviews are now everywhere, but the pioneer of this category is still active: The-PR-Agent/pr-agent (Qodo's open-source version returned after being taken over by the community) made a comeback on GitHub Explore's trending list this week, with 16.4K stars. The four-piece set of /review, /describe, /improve, and /ask covers GitHub, GitLab, Bitbucket, and Azure DevOps, and can be run locally with pip install pr-agent.

AI 代码审查现在遍地开花,但这个品类的开创者还在活跃:The-PR-Agent/pr-agent(Qodo 开源版由社区接管后回归)本周重回 GitHub Explore 热榜,16.4K 星。/review、/describe、/improve、/ask 四件套覆盖 GitHub、GitLab、Bitbucket 和 Azure DevOps,pip install pr-agent 就能本地跑。

Rust CLI · 并行 Agent 必备

Worktrunk: The most popular Git worktree manager that allows 10 AI agents to work in parallel without conflictsWorktrunk:最流行的 Git worktree 管理器,让 10 个 AI Agent 并行干活不打架

AI Agents can now handle long tasks independently, but what is truly scarce is the ability to "manage 5–10 conversations simultaneously". Worktrunk makes the git worktree experience as seamless as switching branches: the wt switch command creates an isolated working directory and launches Claude directly. It is the recognised infrastructure for parallel Agent workflows.

AI Agent 现在能独立干长活,真正稀缺的是「同时管 5-10 个会话」的能力。Worktrunk 把 git worktree 的体验做到像切分支一样顺手:wt switch 一条命令建好隔离工作目录并直接启动 Claude,是并行 Agent 工作流的公认基础设施。

AI 技能包

AI 技能包AI Skill Packs · Give your AI real abilities, out of the box给 AI 装上具体能力,开箱即用

查看全部 44 条 →
技能包 · 今日 GitHub Trending 第一

Claude BugHunter: The 83-Skill Pack that Turns Claude Code into a Senior Bug HunterClaude BugHunter:把 Claude Code 变成资深漏洞猎手的 83 技能包

Claude BugHunter, which is currently ranked number one on GitHub Trending (with 126,000 stars), is a Claude Code skill pack designed for vulnerability discovery and external red team operations. It comprises 83 skills, 15 slash commands, and vulnerability patterns extracted from 681 publicly disclosed reports (433 of which are annotated with specific source references), covering 24 core vulnerability categories, along with an enterprise identity and infrastructure attack matrix.

今天 GitHub Trending 日榜第一(12.6 万星在读)的 Claude BugHunter 是一套为漏洞挖掘和外部红队工作打造的 Claude Code 技能包:83 个技能、15 个斜杠命令、681 份公开披露报告提炼出的漏洞模式(其中 433 条逐条标注了引用来源),覆盖 24 类核心漏洞,外加企业身份与基础设施攻击矩阵。

技能包 · 7.1 万星 · 去 AI 味

Humanizer: Detects 25 types of AI-generated writing and rewrites AI-sounding text into "human language".Humanizer:检测 25 种 AI 写作痕迹、把 AI 腔改写成「人话」的技能

Text written by AI can be recognised at a glance – "It is worth noting", "In summary", three consecutive parallel structures, excessive antithesis. Humanizer (71,000 stars) is an Agent skill that focuses on text: it detects 25 known AI-generated writing patterns and then rewrites the text without changing its original meaning, making it read like it was written by a human.

AI 写的东西一眼就能认出来——「值得注意的是」「总而言之」、排比三连、过度对仗。Humanizer(7.1 万星)是一个专注于文字的 Agent 技能:检测 25 种已知的 AI 生成写作模式,然后在不改变原意的前提下重写,让文字读起来像人写的。

11 个创业体检技能 · 免费无套路

Founder Skill: Get Claude to play Hormozi, Thiel and Jobs, and give your business a start-up health check.Founder Skill:让 Claude 扮演 Hormozi、Thiel 和 Jobs,给你的生意做一次创业体检

The most expensive lesson in entrepreneurship is "finishing a product only to find that no one wants to buy it". Founder Skill is a free Claude skill pack with 4.2K stars: 11 skills that bring a board of directors (trained using the frameworks of Hormozi, Thiel and Jobs respectively), a marketing director, a CFO and a consumer jury consisting of 100 buyer agents to your device, allowing you to run through your business before you spend any real money.

创业最贵的学费是「做完才发现没人买」。Founder Skill 是一套 4.2K 星的免费 Claude 技能包:11 个技能把董事会(分别用 Hormozi、Thiel 和 Jobs 的框架训练)、营销总监、CFO 和一个 100 个买家 Agent 组成的消费者陪审团搬进终端,在你花真金白银之前把生意先盘一遍。

简街官方 · 苏格拉底式私教

Jane Street's first official skill pack: socratic-tutor turns large language models into ruthless Socratic tutorsJane Street 首个官方技能包:socratic-tutor 把大模型变成不讲情面的苏格拉底私教

Quantitative giant Jane Street has publicly released its internal skills repository for the first time. The first skill, socratic-tutor, prevents AI from giving direct answers. Instead, it first conducts a baseline assessment, maps out a skill level from 0 to 5, asks one question at a time, schedules spaced repetition, and finally conducts a final assessment. Within a week of being posted, it received over 10,000 stars.

量化巨头 Jane Street 首次公开内部技能仓库,首发技能 socratic-tutor 让 AI 不再直接给答案,而是先做摸底诊断、画出 0-5 级技能地图、一次只问一个问题、安排间隔复习、最后来场结业考核——上墙一周即破万星。

正能量公益广告今日正能量保持好奇,持续探索,未来的你会感谢今天的开始。挑一个好工具上手 →广告
免费下载专区

免费下载专区真实文件,点击即下,无需注册

全部 168 个资源 →
AI插件1.23MB

dsh-our-free-model:在 DeepSeek Harness 里免登录免费用 DeepSeek V4.1 Flash、Kimi K3

dsh-our-free-model v2.1.0 官方 npm 包

AI工具19.58MB

CC Switch:一键切换 Claude Code、Codex 等 10+ AI 编程工具的 API 供应商

CC Switch v4.0.36 Windows 便携版(免安装)

AI技能包2.01MB

Claude BugHunter:把 Claude Code 变成资深漏洞猎手的 83 技能包

Claude BugHunter v2.1 完整技能包(GitHub 仓库 zip)

AI技能包约 33 KB

Humanizer:检测 25 种 AI 写作痕迹、把 AI 腔改写成「人话」的技能

Humanizer v4.0 技能包(GitHub 仓库 zip)

2.01MB

手把手教你装好 Claude BugHunter:从零跑通第一次授权漏洞测试

Claude BugHunter v2.1 完整技能包(本教程配套)

19.58MB

手把手教你用 CC Switch:10 分钟管好所有 AI 编程工具的供应商与配置

CC Switch v4.0.36 Windows 便携版(本教程配套)

AI插件1.02MB

BLA:给软件穿上「防 AI 逆向」护甲的开源蓝方防御 MCP 插件

BLA(bla-agents)v3.5.1 官方 npm 包

AI工具约 704 KB

Agent Autoloop:让命令行 Agent 后台自己接活,拿不准才来找你拍板

Agent Autoloop v1.9.3 完整源码包(含内核与全部插件)

编辑精选值得先看的几篇

全部内容 →
开源插件 · 免费用上前沿模型

dsh-our-free-model: Use DeepSeek V4.1 Flash and Kimi K3 for free without logging in to DeepSeek Harnessdsh-our-free-model:在 DeepSeek Harness 里免登录免费用 DeepSeek V4.1 Flash、Kimi K3

This is one of the AI projects that has gained the most stars in the last 7 days (No. 1 on the GitHub Trending daily list, with a total of 81,000 stars): Simply install a plugin in DeepSeek Harness (dsh), and you can directly use cutting-edge models such as DeepSeek V4.1 Flash and Kimi K3, without needing to log in, register, or enter any API key. It's completely free and unlimited. Pure plugin installation: No kernel modifications, no build steps, zero dependencies.

这是最近 7 天涨星最快的 AI 项目之一(GitHub Trending 日榜第一,累计 8.1 万星):只需往 DeepSeek Harness(dsh)里装一个插件,无需登录、注册或填任何 API Key,就能直接用上 DeepSeek V4.1 Flash、Kimi K3 等前沿模型,完全免费、不限量。纯插件挂载:不改内核、无构建步骤、零依赖。

开发者工具 · 桌面端 · 15.6 万星

CC Switch: One-click switching of API providers for 10+ AI programming tools such as Claude Code and CodexCC Switch:一键切换 Claude Code、Codex 等 10+ AI 编程工具的 API 供应商

Who among those who use Claude Code, Codex, and Gemini CLI at the same time hasn't experienced the agony of spending an entire evening manually editing config.json? CC Switch (156,000 stars, built with Tauri 2) is a cross-platform desktop manager: Switch the API providers for all AI programming CLIs with one click, and centrally manage their MCP servers, Skills, and Prompts – no more manual editing of configuration files.

同时用 Claude Code、Codex、Gemini CLI 的人,谁没经历过手改一晚上 config.json 的痛苦?CC Switch(15.6 万星,Tauri 2 构建)是一个跨平台桌面管理器:一键切换所有 AI 编程 CLI 的 API 供应商,还能集中管理它们的 MCP 服务器、Skills 和 Prompts,不用再手动编辑任何配置文件。

技能包 · 今日 GitHub Trending 第一

Claude BugHunter: The 83-Skill Pack that Turns Claude Code into a Senior Bug HunterClaude BugHunter:把 Claude Code 变成资深漏洞猎手的 83 技能包

Claude BugHunter, which is currently ranked number one on GitHub Trending (with 126,000 stars), is a Claude Code skill pack designed for vulnerability discovery and external red team operations. It comprises 83 skills, 15 slash commands, and vulnerability patterns extracted from 681 publicly disclosed reports (433 of which are annotated with specific source references), covering 24 core vulnerability categories, along with an enterprise identity and infrastructure attack matrix.

今天 GitHub Trending 日榜第一(12.6 万星在读)的 Claude BugHunter 是一套为漏洞挖掘和外部红队工作打造的 Claude Code 技能包:83 个技能、15 个斜杠命令、681 份公开披露报告提炼出的漏洞模式(其中 433 条逐条标注了引用来源),覆盖 24 类核心漏洞,外加企业身份与基础设施攻击矩阵。

技能包 · 7.1 万星 · 去 AI 味

Humanizer: Detects 25 types of AI-generated writing and rewrites AI-sounding text into "human language".Humanizer:检测 25 种 AI 写作痕迹、把 AI 腔改写成「人话」的技能

Text written by AI can be recognised at a glance – "It is worth noting", "In summary", three consecutive parallel structures, excessive antithesis. Humanizer (71,000 stars) is an Agent skill that focuses on text: it detects 25 known AI-generated writing patterns and then rewrites the text without changing its original meaning, making it read like it was written by a human.

AI 写的东西一眼就能认出来——「值得注意的是」「总而言之」、排比三连、过度对仗。Humanizer(7.1 万星)是一个专注于文字的 Agent 技能:检测 25 种已知的 AI 生成写作模式,然后在不改变原意的前提下重写,让文字读起来像人写的。

图文教程 · 手把手

A Step-by-Step Guide to Setting Up Claude BugHunter: Complete Your First Authorised Vulnerability Test from Scratch手把手教你装好 Claude BugHunter:从零跑通第一次授权漏洞测试

In our previous article, we introduced the Claude BugHunter skill pack, which has risen to the top of GitHub Trending today. In this article, we will take you through the entire process, from downloading and installing the pack to conducting real-world practice on a live range: Equip yourself with 83 vulnerability hunting skills and complete your first authorised penetration test on DVWA and OWASP Juice Shop.

上一篇我们介绍了今天冲上 GitHub Trending 第一的 Claude BugHunter 技能包,这篇带你从下载、安装到真实靶场实战完整走一遍:装好 83 个漏洞猎杀技能,在 DVWA 和 OWASP Juice Shop 上完成你的第一次授权渗透测试。

图文教程 · 手把手

A Step-by-Step Guide to Using CC Switch: Manage All AI Programming Tool Providers and Configurations in 10 Minutes手把手教你用 CC Switch:10 分钟管好所有 AI 编程工具的供应商与配置

In the previous article, we introduced the CC Switch desktop manager, which has received 156,000 stars. This article will guide you from downloading and installing the software to completing your first provider switch, and then to managing the MCP server and Skills in a unified manner. We will use the official portable version that was released today throughout the process. We will use Windows as an example, but the process is the same for macOS/Linux.

上一篇我们介绍了 15.6 万星的 CC Switch 桌面管理器,这篇带你从下载安装到完成第一次供应商切换,再把 MCP 服务器和 Skills 也统一管起来。全程用今天入库的官方便携版,Windows 为例,macOS/Linux 流程相同。

开源 AI 安全 · 蓝方防御 MCP

BLA: An open-source Blue Team defence MCP plugin that equips software with "AI reverse engineering protection" armourBLA:给软件穿上「防 AI 逆向」护甲的开源蓝方防御 MCP 插件

An AI Reverse Engineering Agent (REA) can tear apart a release package in just a few minutes, while the defenders are still using obfuscators from the last century. BLA (Blind LLM Agents) is an open-source blue team defence MCP & CLI that rose to 5th place on the GitHub Trending daily list this week: 31 defence tools covering 26 product formats, using the three-pronged approach of "poisoning → exhaustion → hallucination" to bankrupt AI reverse engineering in terms of computing power.

AI 逆向 Agent(REA)几分钟就能把一个发布包拆个底朝天,防御方却还在用上世纪的混淆器。BLA(Blind LLM Agents)是本周冲上 GitHub Trending 日榜第 5 的开源蓝方防御 MCP & CLI:31 个防御工具、覆盖 26 种产物格式,用「投毒 → 耗尽 → 幻觉」三板斧让 AI 逆向在算力上破产。

国产开源 · Agent 自动化中台

Agent Autoloop: Let the command-line Agent handle tasks in the background on its own, and only come to you for decisions when it's unsure.Agent Autoloop:让命令行 Agent 后台自己接活,拿不准才来找你拍板

The biggest waste of a command-line agent is having someone man the terminal: if you walk away, it stops. Agent Autoloop is a Chinese open-source project that was only created on 3 November and reached 5,909 stars in just two weeks – an automation middleware with zero external dependencies. It turns event sources, agents, and notification channels into pluggable plugins. Tasks queue up and follow up on their own, and only when a decision is needed will a card be pushed to Lark, DingTalk, WeChat Work, or Slack to notify you.

命令行 Agent 最大的浪费是人守着终端:你走开,它就停。Agent Autoloop 是 11 月 3 日才创建、两周冲到 5909 星的国产开源项目——一个零外部依赖的自动化中台,把事件来源、Agent、通知通道全部做成可插拔插件,任务自己排队跟进,只有需要拍板时才把卡片推到飞书、钉钉、企微或 Slack 找你。

开源鼻祖 · AI 代码审查

PR-Agent: The pioneer of open-source AI code review is back, offering a health check for any PR with a single command.PR-Agent:开源 AI 代码审查的鼻祖回归,一行命令给任意 PR 做体检

AI code reviews are now everywhere, but the pioneer of this category is still active: The-PR-Agent/pr-agent (Qodo's open-source version returned after being taken over by the community) made a comeback on GitHub Explore's trending list this week, with 16.4K stars. The four-piece set of /review, /describe, /improve, and /ask covers GitHub, GitLab, Bitbucket, and Azure DevOps, and can be run locally with pip install pr-agent.

AI 代码审查现在遍地开花,但这个品类的开创者还在活跃:The-PR-Agent/pr-agent(Qodo 开源版由社区接管后回归)本周重回 GitHub Explore 热榜,16.4K 星。/review、/describe、/improve、/ask 四件套覆盖 GitHub、GitLab、Bitbucket 和 Azure DevOps,pip install pr-agent 就能本地跑。

11 个创业体检技能 · 免费无套路

Founder Skill: Get Claude to play Hormozi, Thiel and Jobs, and give your business a start-up health check.Founder Skill:让 Claude 扮演 Hormozi、Thiel 和 Jobs,给你的生意做一次创业体检

The most expensive lesson in entrepreneurship is "finishing a product only to find that no one wants to buy it". Founder Skill is a free Claude skill pack with 4.2K stars: 11 skills that bring a board of directors (trained using the frameworks of Hormozi, Thiel and Jobs respectively), a marketing director, a CFO and a consumer jury consisting of 100 buyer agents to your device, allowing you to run through your business before you spend any real money.

创业最贵的学费是「做完才发现没人买」。Founder Skill 是一套 4.2K 星的免费 Claude 技能包:11 个技能把董事会(分别用 Hormozi、Thiel 和 Jobs 的框架训练)、营销总监、CFO 和一个 100 个买家 Agent 组成的消费者陪审团搬进终端,在你花真金白银之前把生意先盘一遍。

图文教程 · 手把手

A Step-by-Step Guide to Setting Up Agent Autoloop: Let Claude Code Handle Tasks in the Background手把手教你装好 Agent Autoloop:让 Claude Code 在后台自己接活

In the previous article, we introduced Agent Autoloop, the "Agent Automation Mid-End". In this article, we'll guide you through setting it up from scratch and going through the complete closed loop of "Assign Task → Run in the Background → Approve Card → Complete Task". The official source code package released today will be used throughout the process, and it is compatible with Windows, macOS, and Linux.

上一篇我们介绍了 Agent Autoloop 这台「Agent 自动化中台」,这篇带你从零装好它,并真实走一遍「派活 → 后台跑 → 卡片拍板 → 干完收工」的完整闭环。全程用今天入库的官方源码包,Windows、macOS、Linux 通用。

图文教程 · 手把手

A step-by-step guide to reviewing code locally with PR-Agent: No need to install a bot – generate a report with a single command手把手用 PR-Agent 本地审查代码:不装机器人,一条命令出报告

Teams may not be willing to install robots in their repositories, but everyone wants to merge good code. This tutorial uses the official PyPI package of PR-Agent v0.50.0, which was added to the repository today, to guide you through the four steps of "installing the package → configuring the key → reviewing a PR locally → understanding the report", without touching any platform configuration throughout the process.

团队未必愿意给仓库装机器人,但每个人都想把好代码关。这篇教程用今天入库的 PR-Agent v0.50.0 官方 PyPI 包,带你走通「装包 → 配 key → 本地审一个 PR → 读懂报告」四步,全程不碰任何平台配置。

谷歌官方 · 设计评审工作流

Elephant-Goldfish Model: Google's open-source "Elephant Drafting, Goldfish Reviewing" specification design plugin, 21.1K starsElephant-Goldfish Model:谷歌开源「大象起草、金鱼评审」规格设计插件,21.1K 星

The biggest pitfall when writing software specifications is "reviewing your own work": the context is too saturated, and blind spots cannot be seen. EGM uses the "Elephant" role to hold all the project context and draft the specifications, and then uses the "Goldfish" reviewer, who has zero context, to pick out flaws with fresh eyes and identify unstated assumptions and oversights. Supports mainstream coding agents such as Claude Code, Codex, and Antigravity.

写软件规格时最大的坑是「自己审自己」:上下文太满,盲区看不见。EGM 用「大象」角色握住全部项目上下文起草规格,再用零上下文的「金鱼」评审员带着新鲜眼光挑毛病,把没说出口的假设和遗漏揪出来。支持 Claude Code、Codex、Antigravity 等主流编码 Agent。

Rust CLI · 并行 Agent 必备

Worktrunk: The most popular Git worktree manager that allows 10 AI agents to work in parallel without conflictsWorktrunk:最流行的 Git worktree 管理器,让 10 个 AI Agent 并行干活不打架

AI Agents can now handle long tasks independently, but what is truly scarce is the ability to "manage 5–10 conversations simultaneously". Worktrunk makes the git worktree experience as seamless as switching branches: the wt switch command creates an isolated working directory and launches Claude directly. It is the recognised infrastructure for parallel Agent workflows.

AI Agent 现在能独立干长活,真正稀缺的是「同时管 5-10 个会话」的能力。Worktrunk 把 git worktree 的体验做到像切分支一样顺手:wt switch 一条命令建好隔离工作目录并直接启动 Claude,是并行 Agent 工作流的公认基础设施。

简街官方 · 苏格拉底式私教

Jane Street's first official skill pack: socratic-tutor turns large language models into ruthless Socratic tutorsJane Street 首个官方技能包:socratic-tutor 把大模型变成不讲情面的苏格拉底私教

Quantitative giant Jane Street has publicly released its internal skills repository for the first time. The first skill, socratic-tutor, prevents AI from giving direct answers. Instead, it first conducts a baseline assessment, maps out a skill level from 0 to 5, asks one question at a time, schedules spaced repetition, and finally conducts a final assessment. Within a week of being posted, it received over 10,000 stars.

量化巨头 Jane Street 首次公开内部技能仓库,首发技能 socratic-tutor 让 AI 不再直接给答案,而是先做摸底诊断、画出 0-5 级技能地图、一次只问一个问题、安排间隔复习、最后来场结业考核——上墙一周即破万星。

插件介绍 · 附完整插件包

Claude-Obsidian: Turn Claude Code and Obsidian into a "local-first" knowledge work partnership. Any information you put in becomes a note with a reference link.claude-obsidian:让 Claude Code 和 Obsidian 变成一对「本地优先」知识工作搭档,任何资料丢进去都变成带引用链接的笔记

Do you have two hundred unread articles piled up in your favourites and dozens of PDFs sitting in your downloads folder, and you want AI to help you process them but are always worried that it will make up sources? This open-source project with over 16,400 stars directly connects Claude Code (or any host compatible with Agent Skills) to the Obsidian vault: you feed it any source, and it generates interlinked Markdown notes with each conclusion citing its source. When you ask questions, the answers are strictly based on the evidence already in the vault, and if it can't find an answer, it simply says so. It runs entirely locally, and your data never leaves your computer.

收藏夹里堆了两百篇没读的文章、下载目录里躺着几十份 PDF,想让 AI 帮你消化却总担心它瞎编出处?这个 16,400+ 星的开源项目把 Claude Code(或任何兼容 Agent Skills 的宿主)和 Obsidian 库直接接通:你把任何来源丢进去,它产出互相链接、每个结论都标明出处的 Markdown 笔记;提问时回答也严格基于库里已有证据,不给答案就直说找不到。全程本地运行,数据不出你的电脑。

MIT 许可 · 16,400+ 星 · v3.0.1 · 本地优先查看 →
正能量公益广告今日正能量学一点,用一点;今天种下的种子,会长成明天的能力。去免费下载专区 →广告