周报 · 2026 年第 44 周
GitHub 新星 · 汇总该周新收录项目
当周新增收录 259 个项目
覆盖日期：2026-10-26 至 2026-11-01 · 北京时间
发布时间：2026/11/2 10:00:16
部分来源已不可公开，本期文章将按其余项目修订后重新展示。下方为已保存的项目资料。
本期项目资料
收录数按本期首次达标收录的项目统计。
- designfuchen/word-atlasTypeScript · 3,714 Stars
雅思背词、间隔复习、词库下载
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
词光 · Word Atlas — Illustrated IELTS vocabulary with 7,298 words, UK/US audio and FSRS review. Open-source app + CC BY 4.0 dataset.
编报时的 AI 中文解读
词光是一款以插画和语境为核心的雅思背词应用，提供7,298个词条、原创插画、英美发音和FSRS间隔复习，无需账号，学习进度保存在本地设备上。
编报时的 README 依据
[](https://x.com/yangleiocx3) <br> <a href="https://0xlei.tech/words/"><img src="https://0xlei.tech/words/beta/banner.png" alt="词光 iOS 首批内测：微信扫码加入内测群" width="100%"></a> **词光 iOS 正在开发，首批内测招募中。** 微信扫上图二维码进群；二维码过期或群满了，打开 [网站](https://0xlei.tech/words/) 点顶栏「内测」，加作者微信拉你进群。 </div> --- 词光是一款以插画和语境为核心的雅思背词应用。看画面理解词义，听发音建立联系，再用主动回忆和间隔复习巩固记忆。无需账号，学习进度保存在自己的设备上。 > **版本说明**：这个仓库提供词光的开源基础版及配套数据集。在线网站独立迭代，功能与仓库可能不同；仓库里的代码和内容分别遵循下方的 MIT / CC BY 4.0 授权。
git clone --depth 1 https://github.com/designfuchen/word-atlas.git cd word-atlas npm ci npm run dev ``` 打开 **http://localhost:5188**。 ```bash npm test # 单元测试 npm run check:atlas # 词库结构与媒体校验 npm run build # 构建到 dist/ ``` 子路径部署：`ATLAS_BASE=/words/ npm run build`。完整说明见 [部署文档](docs/deployment.md)。 **技术栈**：React 19 · TypeScript · Vite · ts-fsrs · WebGL2 ## 开放数据 不仅可以直接使用应用，也可以把词库用于自己的学习工具、Anki 牌组或研究项目。 | 资源 | 开源版内容 | 入口 | | :--- | :--- | :--- | | 词条 | **7,298** 条，CEFR A1–C2、19 个话题 | [JSON](data/word-atlas.json) · [CSV](data/word-atlas.csv) |
| 🌧 **6 段环境声** | [雨天图书馆](https://0xlei.tech/words/sfx/amb-rain.mp3) · [清晨海岸](https://0xlei.tech/words/sfx/amb-sea.mp3) · [林间早晨](https://0xlei.tech/words/sfx/amb-forest.mp3) · [夜晚壁炉](https://0xlei.tech/words/sfx/amb-fire.mp3) · [远航船舱](https://0xlei.tech/words/sfx/amb-voyage.mp3) · [安静咖啡馆](https://0xlei.tech/words/sfx/amb-cafe.mp3) | | ✨ **界面音效** | 答对、答错、翻卡、打字、连胜、点亮星星……每个操作都有自己的声音；读单词时背景音会自动压低 | 全部素材都在仓库里：插画在 `public/atlas/img/`，录音在 `public/atlas/audio/`，这两类按 CC BY 4.0 开放；界面声音和配乐在 `public/sfx/`，随应用一起提供。 </details> ## 本地运行 需要 **Node.js 20+** 和 npm。仓库含完整图片与录音，首次下载可能较久。 ```bash
释义、例句、插画和录音为本项目生成：文字与插画使用生成式 AI，语音使用 ElevenLabs。内容经过自动结构与一致性校验，**尚未逐条人工审校**，可能出现释义、例句、插画或发音错误。它不是官方雅思词表，也不承诺提分。 [内容生成流程](docs/pipeline.md) · [记忆模型](docs/learning-model.md) · [生成脚本](scripts/atlas/) ## 参与贡献 发现问题时，请附上**词条 ID、复现步骤和预期结果**。内容纠错如有词典或其他可靠依据，也请一并提供。 - [报告问题](https://github.com/designfuchen/word-atlas/issues/new/choose)：词条纠错、界面问题或使用建议。 - [贡献指南](CONTRIBUTING.md)：数据同步、代码规范和提交前检查。 - [安全政策](SECURITY.md)：安全问题的反馈方式。 ## 授权 | 范围 | 许可 | | :--- | :--- | | 应用代码 | [MIT](LICENSE) | | 词库文本、插画与词条录音 | [CC BY 4.0](CONTENT_LICENSE.md)，署名 Word Atlas contributors（词光） | | 第三方依赖与字体 | 遵循各自许可，详见相应文件 |
- JoasASantos/REA-PromptsBatchfile · 3,410 Stars
逆向工程提示词、软件分析
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
REAP - Prompts for you to use REA (Reverse Engineer Anything), get the most out of it, and clone or break down any program
编报时的 AI 中文解读
为REA逆向工程工具集提供一套精选提示词，帮助AI编码代理分析软件工作原理，涵盖从原生二进制、固件、内核到游戏、Electron应用、APK和网站等多种目标。
编报时的 README 依据
# REA Investigation Prompts (REAP) — Reverse Engineering with your Coding Agent A curated collection of prompts for [REA (Reverse Engineer Anything)](https://rea.tools/) — the toolset that lets your AI coding agent inspect how software works, from native binaries, firmware and kernels to games, Electron apps, APKs and websites. Prompts are ready to paste into your agent. Replace everything in `<angle brackets>` and keep each request focused on **one feature** — narrow questions get evidence-backed answers. ## Star History
## Install as agent skills (recommended) `setup.sh` (macOS/Linux) or `setup.bat` (Windows) copies every prompt section into your coding agent's skills folder (`~/.hermes/skills`, `~/.claude/skills` or `~/.codex/skills`): ```bash ./setup.sh # auto-detect installed agents ./setup.sh hermes # or: claude-code | codex | all ``` ```bat setup.bat :: Windows ``` What you get:
Every prompt asks the agent to separate *confirmed facts* from *hypotheses* and to cite evidence — treat that as non-negotiable. --- ## 0. Install & verify ```text Install REA and connect it to this agent using: npx rea-agents@latest setup First, show me the installation plan and the configuration changes for my approval. After the install, verify the connection and list the available tools and workflows. ``` --- ## 1. Getting started ### First exercise — trace a CSV export ```text Download https://rea.tools/examples/notes-example.zip into a new folder and extract the contents.
| — | [Wildcard prompt](#wildcard-prompt-start-here) · [Prompt-writing tips](#prompt-writing-tips-why-these-work) · [Next steps](#next-steps-see-what-people-built-with-rea) | > ⚠️ **Use responsibly.** Only analyze software you own or are explicitly authorized to test. Anti-cheat, DMA, DRM, kernel and vulnerability prompts below are for **security research, interoperability, CTFs, bug bounties and testing the protection of your own software**. Bypassing protection of software you don't own, cheating in online games, or weaponizing a found vulnerability may violate law and terms of service.
- deepseek-ai/dsh-superagent-prototypeTypeScript · 7,860 Stars
代码任务自动化
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
Towards a More Autonomous Agent System
编报时的 AI 中文解读
dsh-superagent 是一个面向单个 GitHub 仓库的主动编码代理，集成在 DeepSeek Harness 中，可监听 GitHub 和飞书，自动处理工作项并生成拉取请求。
编报时的 README 依据
dsh-superagent is a proactive coding agent for one GitHub repository, packaged as a bundle for [DeepSeek Harness](https://www.deepseek.com/en/harness/). It listens to GitHub and, optionally, Feishu. When someone allowed asks the agent for work, it opens a work item, works on it in its own agent session and directory, and answers in the original conversation, for example with a pull request. A web panel shows what needs a person.
# dsh-superagent > **Research prototype for reference only, not an official DeepSeek product.** This code grew out of our team's day-to-day development work, with much of the engineering done by coding agents. Treat it as research code, not production-ready software. It may still contain bugs, and its coding style and engineering practices may be inconsistent. Review and adapt it before using it in your own environment. >
- heydamir/unslopJavaScript · 4,918 Stars
UI优化
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
A Claude skill that turns vibecoded, AI-looking UI into a designed interface
编报时的 AI 中文解读
这是一个Claude技能，用于将AI生成或粗糙的UI界面转化为有设计感的界面，通过审计、设定方向、逐步修复的流程，适用于多种前端技术栈。
编报时的 README 依据
For Claude Code, copy the skill into your skills folder. ```bash git clone https://github.com/heydamir/unslop mkdir -p ~/.claude/skills && cp -r unslop/skills/unslop ~/.claude/skills/ ``` For a single project, copy it into `.claude/skills/` inside that project instead. ## Use In Claude Code, type `/unslop` with what you want, or just ask in your own words and Claude picks the skill up by itself. ``` /unslop this dashboard /unslop the hero of https://example.com /unslop my landing page and make the story video so I can post it ```
- **Visuals.** Shows your product instead of a glossy blob. When a hero needs a scene or an object, it can generate one through an image service like fal, Replicate, OpenAI images or Higgsfield. If none is connected, it offers to connect one as an MCP server, and it shows the prompt and asks before every paid generation. Dashboards and apps rarely need one. - **Story video.** A square video under 10 seconds of your screen unslopping itself, ready to post. It needs Playwright and ffmpeg. Try it on the two vibecoded screens in [examples](examples). ## Hire the unslopper
- ccch1mneyyy/dsh-GoDart · 4,722 Stars
本机模型运行、远程连接、聊天辅助
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
dsh-go — DeepSeek Harness 装进口袋：手机本机运行，也能遥控你的电脑。
编报时的 AI 中文解读
DSH Go 是 DeepSeek Harness 的 Android 客户端，可让手机本机运行 DSH，也可远程连接电脑上的 DSH。适合希望在移动端使用和操控 DSH 的用户。
编报时的 README 依据
<picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/pet-girl-sleeping-zh-dark.gif"><img src="docs/assets/pet-girl-sleeping-zh-light.gif" width="168" alt="鲸娘 睡觉"></picture> </p> <br> <a id="download"></a> <h3 align="center">下载 DSH Go</h3> <p align="center">从 GitHub Releases 或官网下载页获取最新版 APK，常见手机选 arm64 包。</p> <p align="center">
- **手机操控**：在你开启权限后，让本机引擎读取屏幕、点按和输入，并可随时停止。 - **分享与外观**：分享、导出和回放对话，调整主题、图标和桌宠。 功能随主机和连接方式不同。DSH Remote 当前只支持图片附件，不能在 App 内打开网页版，插件管理限于开关，且不支持 DS Bot 和对话日志导出。详见[本机引擎与远程主机](docs/guide/zh/03-engine-hosts.md)。 ## 安装 安装要求为 Android 8.0（API 26）及以上。Release 页面上的 **arm64** 安装包适用于常见的 64 位 Android 手机；x86_64 设备和模拟器用的 x64 包在发布说明末尾的「其他版本下载」里。仓库没有 32 位 ARM 构建。 1. 从 [GitHub Releases](https://github.com/ccch1mneyyy/dsh-go/releases/latest) 或[官网下载页](https://go.dshtui.com/download)下载对应架构的 APK，并按系统提示允许该安装来源。 2. 首次启动时，选择在本机安装运行环境，或连接远程主机。 3. 为所选主机配置模型和所需凭据，再开始对话；部分远程渠道需要在主机侧配置。
<p align="center">和官方网页版一键切换，停在同一段对话，下次打开还是你上次用的那个。</p> <p align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="docs/assets/f-web-zh-dark.webp"> <img src="docs/assets/f-web-zh-light.webp" width="100%" alt="同一段对话的原生界面与网页版"> </picture> </p> <br> <h3 align="center">分享卡、对话回放和链接</h3> <p align="center">回复一键做成分享卡。整段对话生成链接，对方在浏览器里回放，链接端到端加密。</p> <p align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="docs/assets/f-share-zh-dark.webp"> <img src="docs/assets/f-share-zh-light.webp" width="100%" alt="分享卡、对话回放、我的分享链接">
Android 8 及以上。手机直接下载 Release 页面上的安装包即可，模拟器用的 x86_64 版在发布说明末尾的「其他版本下载」里。在手机上运行 DSH 要先准备内置的运行环境，请预留约 1 GB 空间；只连电脑或服务器上的 DSH 可以不装。 </details> <details> <summary><b>开源吗？</b></summary> <br> 开源，代码就在这个仓库，以 [AGPL-3.0](LICENSE) 发布，`sdk/` 和 `examples/` 为 MIT。DSH Go 的名称和标识，以及 Deepy、鲸娘两位吉祥物，不在开源授权范围内，见 [ASSETS.md](ASSETS.md)。 </details> ## 从源码构建 以下流程面向 **Linux x86_64**。当前工具链脚本默认安装 JDK 17、Flutter 3.47.7 和 Android SDK/NDK；App 要求 Dart `^3.13.5`。准备 Bash、Git、curl、tar、unzip、xz、Python 3、zstd（Python 未提供 `compression.zstd` 时用于解包 zstd 压缩的 deb）、file 和常见 GNU 命令；重新解析 Ubuntu 软件包锁时还需要 gpgv。
仅连接远程主机时，不必在手机上安装本机运行环境。手机操控只对正常模式下的本机引擎开放，需要按界面说明开启无障碍等权限；不打扰模式另外要求 Android 14 及以上，以及 Shizuku 或自连 ADB。首次使用可按[使用指南](docs/guide/README.md)逐项配置。 本地引擎默认使用 PRoot 兼容启动。开发者选项中的直接启动默认关闭，开启后的兼容性和回退方式见[常见问题](docs/guide/zh/11-faq.md)。当前 Android 构建以 APK 侧载为目标，不是 Google Play 发布配置。 ## 数据与权限 模型请求会发送给你配置的提供方。开启手机操控后，任务需要的屏幕文字可能发送给模型，支持看图的模型也可能收到截图；请先阅读 App 的权限说明。 “发送匿名使用统计”默认开启，可在首次引导或“设置 → 隐私与分享”关闭。采集范围和关闭后的处理见[常见问题](docs/guide/zh/11-faq.md)。分享内容前，请检查预览；收到链接或文件的人可以保存其中的内容。 ## 常见问题 <details> <summary><b>和 DeepSeek 官方是什么关系？</b></summary> <br>
- interplains/Zero-Breach-ToolkitPython · 4,157 Stars
间谍软件检测
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
Free, open-source check for traces of mercenary spyware (Pegasus, Predator, Graphite, Candiru and more) on computers, Android phones and iPhone backups. Double-click apps for Windows, Mac and Linux, in 27 languages.
编报时的 AI 中文解读
免费开源工具，用于在电脑、Android手机和iPhone备份中检测已知商业间谍软件（如Pegasus、Predator等）的痕迹。提供Windows、Mac和Linux的图形界面，支持27种语言，无需技术经验即可使用。
编报时的 README 依据
5. Close your web browsers and chat apps (Chrome, Edge, WhatsApp, Signal…). 6. Open the **Zero-Breach-Toolkit-main** folder (if there is another folder with the same name inside, open that one too), then the **scan** folder, and double-click **run-scan.cmd**. If Windows shows a warning, click **Run** (or **More info**, then **Run anyway**). 7. A menu appears. Type the number next to **Update the indicator lists** and press Enter, then type `y` and press Enter. When it finishes, press Enter. 8. Press Enter again to start **This computer - standard check**. It takes about 15 minutes.
3. Open **System Settings → Privacy & Security → Full Disk Access**, click **+**, choose **Applications → Utilities → Terminal**, and turn it on. Without this, macOS hides Safari's records from the check. **Run the check** 4. Close your web browsers and chat apps. 5. Open **Terminal** (press Cmd+Space, type `Terminal`, press Enter). Copy this line, paste it into Terminal and press Enter: ``` cd ~/Downloads/Zero-Breach-Toolkit-main && python3 -X utf8 scan/zero-breach-toolkit.py ``` 6. A menu appears. Type the number next to **Update the indicator lists** and press Enter,
the [Mobile Verification Toolkit](https://github.com/mvt-project/mvt-indicators) project, the Citizen Lab and Echap's stalkerware research, and a hand-curated set from Citizen Lab reports (Candiru / DevilsTongue and Paragon's Graphite) with every line cited. The curated set works offline; `--update` downloads the rest from a fixed list of official sources. --- ## Install and run You need Python 3.8 or newer ([python.org](https://www.python.org/downloads/); on Windows, tick "Add python.exe to PATH" when installing). No administrator rights and no extra
3. **Recommended: download the latest indicator lists.** This is optional and is the only step that uses the internet; the checks themselves never do. Without it only a small built-in set is used. Repeat it every few weeks. ``` python -X utf8 scan/zero-breach-toolkit.py --update ``` 4. **On a Mac, give Terminal Full Disk Access first** (System Settings → Privacy & Security → Full Disk Access → add Terminal, then reopen it). Without it macOS blocks Safari's records and some folders, and the report lists them as not checked.
# Zero Breach Toolkit A free, open-source check for traces of known mercenary spyware (Pegasus, Predator, Graphite, Candiru and others) on computers, Android phones and iPhone backups. - **Read-only.** It does not modify source evidence or system settings. It writes only temporary database snapshots and the reports you ask for. - **No network requests during a scan.** It never looks up spyware addresses and has no upload feature. Python network calls are blocked and redirected paths are refused. This is not an operating-system network sandbox; see [PRIVACY.md](PRIVACY.md).
- **States its limits.** Every report says what was found, how strong the evidence is, and what could not be checked. A clean result is never presented as proof that you are safe. Made by archivegenocide. Built on [Zero Breach](https://github.com/interplains/zero-breach). > **If you think you are being targeted right now, get expert help first.** You do not > need to run this tool before asking. Journalists, activists, human-rights defenders and > civil society can contact the > [Access Now Digital Security Helpline](https://www.accessnow.org/help/) (free, 24/7,
- microsoft/TokenEconomizerPython · 3,958 Stars
编码代理降本
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
Experience-Driven Black-Box Cost Control for Coding Agents
编报时的 AI 中文解读
TokenEconomizer 是一个面向编码代理（如 Claude Code、Codex 等）的插件，通过从历史会话中学习，利用微技能、工具裁剪和失败循环干预等机制，在不修改代理本身的前提下降低其运行成本。
编报时的 README 依据
Copilot CLI and opencode run the plugin from the clone, so leave it in place. Then start the agent in one of your repositories. **Claude Code** ```sh git clone https://github.com/microsoft/TokenEconomizer.git claude plugin marketplace add "$PWD/TokenEconomizer" claude plugin install tokeneconomizer@tokeneconomizer ``` Start `claude`, trust the folder (**Yes, I trust this folder**), and run `/tokeneconomizer:status`, allowing its script if asked. Once per machine, and again after Claude Code updates, run `/tokeneconomizer:toolprune refresh-catalog`. **Codex** ```sh
full path; `git pull` in the clone updates it. - **opencode** support is experimental: it was tested on 1.19.5 and 2.2.0 only, and not in the 2.x terminal UI. ### Everyday commands Work as usual; the plugin needs nothing from you. To see or change what it does, use its commands. They have the same names on every agent; only the way you type them differs: | Agent | Type a command as | |---|---| | Claude Code, Copilot CLI | `/tokeneconomizer:<command>` | | Codex | `$tokeneconomizer:<command>` | | Pi | `/te <command>` |
- **Keep every action reversible.** A fact that later sessions contradict is dropped, removed tools can be given back, and the loop guard only redirects and tells you: stopping a turn is your call. ## Quick start You need one of these agents, signed in any way it supports, and Python 3.10 or newer on `PATH`. The plugin needs no API key of its own and never handles credentials, so any plan or login works. ### Install Run these commands in the folder where you want to keep the clone. Claude Code, Codex, the
as `./TokenEconomizer` as the name of a GitHub repository. To update: | Agent | Update | |---|---| | Claude Code, Copilot CLI, opencode | `git pull` in the clone | | Codex | `git pull` in the clone, then the `codex plugin add` line again | | Pi | `pi update --extensions` | **[`plugin/README.md`](plugin/README.md)** walks through each agent in more detail, including uninstalling, using Copilot with your own model key, and troubleshooting. ### Supported versions | Agent | Supported | Older versions | |---|---|---| | Claude Code | 2.1.300 to 2.1.310 | not supported |
| Copilot CLI | whole MCP servers | `.github/copilot/settings.local.json` | | Pi | nothing (Pi loads only four tools) | — | | opencode | nothing yet | — | Tool pruning adds only its own entries and never touches deny rules you wrote; a settings file it creates is kept out of git (`.git/info/exclude`). On Claude Code, the session-start line that says what it left out shows only in the detailed transcript (**Ctrl+O**), so check with `/tokeneconomizer:toolprune`. `toolprune restore` removes the plugin's entries and nothing else,
- wangzan101/MikaKotlin · 3,931 Stars
App自动化操作、网页自动化
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
An on-device AI agent for Android that operates apps and the web to get things done for you.
编报时的 AI 中文解读
Mika 是一款运行在 Android 上的 AI 代理应用，能根据用户的一句话指令，在后台屏幕操作 App 和网站，完成诸如发送消息、浏览网页、执行脚本等任务，用户可继续使用手机。
编报时的 README 依据
| <br>后台操作微信，给联系人发消息 | <br>在招聘网站批量打招呼 | | <br>双击胶囊，把这一屏发给 Mika | <br>网页办事：搜索、筛选、读出结果 | --> ## 能做什么 你用一句话交代一件事，Mika 在手机上替你办。 ### 操作任何 App，不占你的屏幕 手机上装的 App 都能操作，不用事先适配：Mika 像人一样看屏幕、找按钮、点击输入，看得到就能操作。 - 在一块单独的后台屏上打开 App、读界面、点击、输入、滚动，你照常用手机；对话页上的小画面实时显示它在做什么。 - 能逐屏读长列表和聊天记录，比如「帮我总结一下某个群今天都聊了些啥」。 - 界面读不出来的 App（自己画界面、控件没有名字）也能用：Mika 用随 APK 自带的文字识别读屏上的字；配了能看图的模型时，还能看图找按钮和输入框。 - 能把一段文字直接分享进某个 App（发消息、记笔记），收件人和发送仍由 Mika 在那一页上点。
- 系统问「用哪个 App 打开」、装了两份的 App 问「打开哪一个」、一个 App 打开另一个 App（比如跳到付款页）时，Mika 认得出来：能处理的接着做，要你选或要你亲手处理的，会停下来请你在屏幕上操作。 - 需要你亲手处理的事（登录、验证码、付款）会发通知请你接管：App 挪到主屏，你处理完交还，Mika 重新核实再继续。 ### 内置浏览器 - 任何网站都能打开：在后台打开网页，用真实的点击和键盘输入操作页面，默认按电脑版网页打开。 - 能读页面正文，也能读页面自己收到的数据；支持下载、上传。 - 登录过的网站保持登录；侧栏「浏览器」里可以自己浏览，在这里登录的网站 Mika 办事时也能直接用。 - 翻页有节奏：同一个网站连着换页会停一停，对限制浏览快慢的网站停得更久；认得出风控拦截页，交给你处理而不是硬闯。 ### 此刻的屏幕 - 你在别的 App 里时，屏幕边上有一颗 Mika 悬浮胶囊，显示后台任务的进展、要你处理的事和结果，点开可以插话、回答、同意或停止。 - **双击胶囊**，把你正在看的这一屏发给 Mika，比如「这个怎么回」「帮我记下来」；也可以在卡片里截屏、补一句话再发。 ### 定时与条件任务 - 「明早 9 点提醒我」「每天早上看看天气，下雨才告诉我」「每周五做某件事」。
<div align="center"> <img src="docs/media/logo.png" alt="Mika" width="120" /> # Mika **你手机上的私人 AI 助理，替你操作 App、上网、把事办成。** 它在一块看不见的后台屏上干活，你照样刷你的手机。 An on-device AI agent for Android that operates apps and the web to get things done for you — on a hidden background screen, so you can keep using your phone. <a href="https://github.com/wangzan101/Mika/releases/latest"><img src="https://img.shields.io/github/v/release/wangzan101/Mika?include_prereleases&label=version" alt="version" /></a> <img src="https://img.shields.io/badge/Android-13%2B-3DDC84" alt="android" />
| 推荐 | 安卓 14 及以上：后台屏有自己的焦点，不影响你在主屏打字；不同对话的任务同时进行也需要它 | | Linux 沙箱 | 仅 arm64 设备；其他设备照常安装，沙箱显示「此设备不支持」 | | 模型 | 一个支持工具调用的大模型 API Key | ### 1. 下载安装 从 [Releases](https://github.com/wangzan101/Mika/releases/latest) 下载最新的 APK 安装。国内下载慢的话，发布说明里有「国内备用下载」网盘链接。 ### 2. 打开无障碍 在系统「无障碍」设置里找到 Mika 并打开。Mika 靠它读界面、做点按；打开后配对无线调试时它还能自己读出配对码。 ### 3. 配对无线调试（只需一次） 打开设置里的「设置手机操作」向导，照着清单做：打开开发者选项、USB 调试、无线调试，然后用 6 位配对码配对。 为什么要这一步：Mika 有一个内置的手机助手，要在后台屏上启动和操作 App，需要借用安卓自带的无线调试来启动它。全程在手机上完成，不需要电脑和数据线。配对过后，如果无线调试被系统关掉了（换 Wi-Fi、重启手机后安卓会关掉它），Mika 在你允许的前提下会自己把它重新打开。 只用上网、文件和聊天的话，这一步可以先跳过。 ### 4. 添加模型
打开设置里的「模型」（第一次打开 App 时也会引导你）。 <!-- generated:providers:start --> <!-- 这一段由 ./gradlew updateReadme 生成，不要手改 --> 内置 DeepSeek、通义千问、Kimi、智谱、OpenRouter、Anthropic、Gemini 等服务商预设；其他服务商选「自定义」，OpenAI 兼容、Anthropic、Gemini 接口都能接。 <!-- generated:providers:end --> 1. 添加服务商：选一个预设，或选「自定义」填接口地址；填 API Key。 2. 点「测试并保存」，连上第一家时会自动选一个主模型。 3. 主模型要支持工具调用。可选：看图模型（看懂界面、看图找按钮）、轻量模型（便宜快的即可，处理判断结局、提取记忆等小事）。 ### 5. 试一试 空对话里有几条示例，点一下就能改了再发。也可以这样说： - 「帮我总结一下微信群今天都聊了些啥」 - 「打开小红书，帮我总结一下某个博主最近发的内容」 - 「帮我订明天早上 9:40、9:50、10:00 的闹钟」 - 「每天早上 8 点看看天气，下雨才告诉我」 ## 询问 / 自主模式 全局一个开关：设置里的「开发者选项」→「自主模式」，关着是询问，打开是自主。 | 模式 | 行为 |
> Status: **public beta**, changing quickly (see [Releases](https://github.com/wangzan101/Mika/releases) for versions and changes). Install on Android 11+, tasks need Android 13+, Android 14+ recommended. The sandbox runs on arm64 devices only. The app UI and docs are in Chinese for now. Licensed under Apache-2.0. ## Beta 说明 - 这是早期公开测试版，功能和界面还在快速变化，升级时可能有不兼容的改动。 - 目前主要在一台安卓 14 手机上测试过。别的安卓版本、别的手机上可能遇到没见过的问题，欢迎[反馈](#参与贡献)。 - Mika 会替你操作第三方 App 和网站，请先读[免责声明](#免责声明)。 <!-- 演示素材录好放进 docs/media/ 后，去掉这段注释的首尾即可显示。 ## 演示 | | | |---|---|
| **网站** | 哔哩哔哩、豆瓣、抖音、Instagram、百度贴吧、微博、搜狗微信、小红书、小宇宙、YouTube、知乎、BOSS直聘 | | **通用做法**（技能） | 把内容整理成文字、处理 PDF 文档、查清一件事、消息与邮件 | <!-- generated:knowledge:end --> ### 成批的事 - 成批做事（比如逐个打招呼、投递）直接开始，不用先跟你确认名单；每两次之间隔开一段随机时间，做到网站自己说「到上限了」就停下告诉你。 ### 过程可查 - 回答下面列出来源（读过的页面、截图）；每次任务都有可回看的过程记录。 - 每次任务一行小结：总用时、和模型来回几次、用了多少 Token、多久看到第一次反馈；设置里的「Token 用量」按天、按服务商、按模型统计。 ### Linux 沙箱 - 每个对话一个工作区，可以跑 Python、Node 和 shell 命令，处理你发来的文件，把结果存到手机的「下载/Mika」。 - 默认断网，需要时在设置里的「文件与计算环境」打开「允许脚本联网」。只支持 arm64 设备。 ## 快速开始 ### 系统要求 | 项目 | 要求 | |---|---| | 安装 | 安卓 11 及以上 | | 执行任务 | 安卓 13 及以上（后台屏和内置手机助手的要求） |
- AienTech/dossierTypeScript · 3,784 Stars
求职管理、简历生成、求职信生成
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
A self-hosted job search app that writes tailored resumes and cover letters from your profile.
编报时的 AI 中文解读
Dossier 是一个自托管的求职应用，它根据你的个人资料为每个职位生成定制简历和求职信，并跟踪从草稿到录用的整个申请流程。适合希望在自有服务器上管理求职过程的个人用户。
编报时的 README 依据
<img alt="Next.js 16" src="https://img.shields.io/badge/Next.js-16-000000?logo=nextdotjs"> <img alt="PostgreSQL 18" src="https://img.shields.io/badge/PostgreSQL-18-4169E1?logo=postgresql&logoColor=white"> </p> Dossier keeps your career history in one profile and turns it into a clean resume. For every job you apply to, it reads the posting, shows how you match, and drafts a tailored resume and cover letter that only use facts from your profile. You refine them in a chat, download them as PDFs and track each application from draft to offer.
It's for one person on a small server: one password, your data in your own PostgreSQL database, and an AI model of your choice. That can be a local model through Ollama or LM Studio, so your profile and documents stay on your own hardware, or a hosted one (any OpenAI-compatible API, Anthropic or Mistral), which then receives what it writes from. [Privacy](https://github.com/AienTech/dossier/wiki/Privacy) says what goes where.
- [Privacy](https://github.com/AienTech/dossier/wiki/Privacy): what stays on your server and what leaves it. [PRIVACY.md](PRIVACY.md) has the details. - [Environment variables](https://github.com/AienTech/dossier/wiki/Environment-variables): every setting the server reads. ## Known limitations - **Desktop only:** made for a laptop or desktop browser, not for phones. - **Single user:** one password and one profile; there are no accounts.
**Get started:** [run it on your computer](#run-it-on-your-computer) in a few minutes, or [self-host it](#self-hosting) on a server. The [installation guide](https://github.com/AienTech/dossier/wiki) covers every way: Docker with or without Compose, Node.js, a plain Ubuntu server, CapRover, Coolify and Dokploy. ## Demo 
All you need is [Docker Desktop](https://docs.docker.com/desktop/) on macOS or Windows, or Docker Engine with its Compose plugin on Linux, and about 2 GB of free memory. Run the commands in a terminal: macOS's Terminal, Linux, or WSL on Windows. **1. Fetch the Compose file** into a folder of its own, and make the secrets Dossier needs: ```bash mkdir ~/dossier && cd ~/dossier curl -fsSL -o compose.yml https://raw.githubusercontent.com/AienTech/dossier/main/compose.production.yml echo "POSTGRES_PASSWORD=$(openssl rand -hex 24)" > .env echo "DOSSIER_SECRET=$(openssl rand -base64 32)" >> .env
To put Dossier on a server, on your own domain with HTTPS, follow the [installation guide](https://github.com/AienTech/dossier/wiki#self-host-on-a-server): a plain Ubuntu server with Docker Compose and Caddy (or Node.js, without Docker), CapRover, Coolify or Dokploy. Give the server at least 2 vCPUs and 2 GB of memory (see [Performance and sizing](#performance-and-sizing)). ## Documentation The [wiki](https://github.com/AienTech/dossier/wiki) has what comes after installing:
- **English interface:** the app itself is in English. Documents can be written in other languages; in the 21 built-in ones, their headings, dates and language levels are in that language too, and in one added with **Other language…** those stay in English. - **One instance:** run a single copy of the app per database. The nightly backups, the AI job queue and the PDF browser live inside the app's process. ## Performance and sizing Measured on 28 September 2026 with the production build in a container limited to 2 vCPUs and 1 GB:
- l0ng-ai/openideaTypeScript · 3,619 Stars
AI辅助编程
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
The IDE you know, rebuilt in Rust. IntelliJ IDEA–style keymap, tool windows and search-everywhere in a fast native app: editor, code insight, debugger, Git, terminal, database and HTTP clients for 60+ languages.
编报时的 AI 中文解读
OpenIDEA 是一个用 Rust 重写的原生 IDE，提供 IntelliJ IDEA 风格的键位、工具窗口和全局搜索，支持 60 多种语言的编辑器、代码洞察、调试器、Git、终端、数据库和 HTTP 客户端，适用于希望获得快速启动和低内存占用的开发者。
编报时的 README 依据
OpenIDEA brings the IntelliJ IDEA way of working — its keymap, its tool windows, its search-everything habits — to a native app with a Rust core. Editor, code insight, debugger, Git, terminal, database and HTTP clients, for 60+ languages. Builds for macOS, Windows and Linux. ## Benchmark Opening [freqtrade](https://github.com/freqtrade/freqtrade) (Python, ~770 files) with warm caches, median of 5 runs: | | OpenIDEA | IntelliJ IDEA 2026.2.3 | |---|---|---| | Window with the project | **0.9 s** | 3.1 s | | Ready (CPU settled) | **1.1 s** | 13.2 s |
Rust 1.96, Node 22, pnpm and the [Tauri prerequisites](https://v2.tauri.app/start/prerequisites/). ## License [Apache 2.0](LICENSE). Bundled third-party assets are listed in [`THIRD_PARTY_LICENSES.md`](THIRD_PARTY_LICENSES.md). <sub>Not affiliated with JetBrains. IntelliJ IDEA is a trademark of JetBrains s.r.o.</sub>
| Memory (USS, all processes) | **169 MB** | 2011 MB | | CPU time until ready | **2.4 s** | 47.9 s | | Idle CPU | **4 %** of a core | 7 % of a core | | Quit | **0.3 s** | 5.9 s | A release build of OpenIDEA from `main` and IDEA with default settings and plugins, on Windows 11, 32-core AMD, 62 GB RAM. IDEA analyses Python more deeply, so "ready" doesn't mean the same work on both sides. Both are measured from the outside, the same way: see [`perf/compare`](perf/compare) to run it on your machine and projects. ## Build ```sh pnpm install pnpm tauri dev ```
- valentderah/totk-pc-recompC · 3,234 Stars
游戏重编译
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
Static recompilation of The Legend of Zelda: Tears of the Kingdom to native Windows x86-64, with a Vulkan renderer. Game files not included.
编报时的 AI 中文解读
将《塞尔达传说：王国之泪》静态重编译为原生 Windows x86-64 应用，使用 Vulkan 渲染器，游戏文件需用户自行提供。适合希望以原生性能在 Windows 上玩该游戏的玩家。
编报时的 README 依据
- Shaders prepared at install instead of during play ## Planned - Linux and the Steam Deck - Mods made for emulators: unpack a mod into a `mods` folder - Custom aspect ratios, including ultrawide - VRR ## Installing Requirements: Windows 10 or later, an x86-64 CPU with AVX2, a GPU with Vulkan 1.3 and `VK_EXT_external_memory_host`, about 36 GiB free during setup (20 GiB after). 1. Download `TotK-Setup-<version>-windows.zip` from [Releases](../../releases/latest) and unpack it. 2. Start `TotK-Setup.exe`. If SmartScreen warns about an unsigned app, choose *More info*, then *Run anyway*.
3. Pick your game files, the language and an install folder, and press *Start*. The setup builds the game once. 4. Play `totk.exe` from the install folder. To update, run a newer `TotK-Setup.exe` on the same folder; saves and settings are kept. Settings are in the in-game menu and in `totk.ini` ([totk.ini.example](totk.ini.example) lists every key). Controls: [docs/controls.md](docs/controls.md). ## Credits TotK Recomp builds on the projects below. What was taken from each, with its license, is listed in [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md).
# TotK Recomp **TotK Recomp** is a static recompilation of *The Legend of Zelda: Tears of the Kingdom* (Nintendo Switch, 2023) into a native Windows x86-64 application ([docs/architecture.md](docs/architecture.md)). > [!IMPORTANT] > **No game files are included.** You need your own legally obtained copy of *Tears of the Kingdom*, dumped from your own console: the base game and the 1.2.1 or 1.4.3 update. Instructions for obtaining those files will not be provided. > > The game is built from your files on your machine, so the result contains the game. It is for your own use. Do not share it.
## Status Early development. The game is playable on Windows; visual bugs and crashes are still expected. Report problems as issues, following [CONTRIBUTING.md](CONTRIBUTING.md). ## Features - Native code, not emulation - Game updates 1.2.1 and 1.4.3, with saves that carry over between them - 60 fps by default; up to 144 fps or the monitor's refresh rate (experimental) - Custom resolutions (1080p, 2k...), supersampling and AMD FSR 1 upscaling - In-game settings menu (F1) and performance overlay (F10) - Keyboard, mouse and gamepad with remapping, rumble and gyro
- SoMarkAI/OXRPython · 2,843 Stars
文档解析
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
An open-source document parsing model with SOTA accuracy and speed, using an RSI loop during post-training.
编报时的 AI 中文解读
OXR 是一个开源文档解析模型，可将 PDF 和文档图片转换为带有布局坐标的 Markdown 或 JSON 数据，适用于需要从文档中提取结构化信息的场景。
编报时的 README 依据
## Quick Start <a id="free-api"></a> ### 0. 🔥 🆓 Free API OXR provides a free API so you can test parsing on your own documents without registering or requesting an API key. We'll keep the free API online until we can no longer afford to run it. ⚠️ Note: We do not recommend using the Free OXR API in production. We do not guarantee service stability, and requests are queued, so processing speed may vary slightly. | Online Demo | https://free-oxr.somark.ai | https://free-oxr.somark.cn | |-|-|-| | API URL | https://free-oxr.somark.ai/api/v1 | https://free-oxr.somark.cn/api/v1 |
modelscope download SoMark/OXR-1.0 --local_dir ./models/OXR-1.0 ``` #### 1.2 Start the Services ##### 1.2.1 Run with Docker (Recommended) We recommend Docker to configure and start all services together. ```Bash docker build -f docker/Dockerfile.allinone -t oxr-1.0 . docker run --rm --gpus all --ipc=host \ -p 127.0.0.1:8000:8000 \ -v "$PWD/models/OXR-1.0:/models/oxr-model:ro" \ oxr-1.0 ``` Once the startup logs show that the services are ready, open http://localhost:8000/demo or http://localhost:8000/docs.
Open http://localhost:8000/demo and upload a PDF or image. You can view the original file, Markdown preview, source, layout bounding boxes, and reading order, and download the Markdown. The demo uses a temporary file cache that expires after 1 hour by default and is renewed during processing. Cached results become inaccessible once the cache expires or is cleared, so download your results promptly. #### 1.4 Send a Parsing Request
somark --base-url https://free-oxr.somark.ai/api/v1 --api-key dummy \ parse document.pdf --formats md,json --out ./parsed-output/ ``` ### 1. Run Locally ```Plain Text Operating system: Linux GPU: NVIDIA (with BF16 support) Driver: NVIDIA driver (compatible with CUDA 13.0) Python: 3.12 (for model downloads / running from source) GPU memory to get started: 24GB on a single GPU (for low-concurrency testing) Recommended GPU memory: ≥48GB on a single GPU (for long documents / higher concurrency) ``` #### 1.1 Download the Code and Models
- InukaAsith/TwinDeskJavaScript · 2,748 Stars
跨设备协同、屏幕共享、文件传输
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
All your computers and your phone, working as one desk. Screens, apps, files, clipboard, calls, texts, notifications, mouse and keyboard, shared between Mac, Windows, Linux and Android. Free and open source. No account, no cloud, and now no shared network needed.
编报时的 AI 中文解读
TwinDesk 是一款免费开源的跨设备协同工具，让 Mac、Windows、Linux 和 Android 设备共享屏幕、应用、文件、剪贴板、鼠标键盘、通知和通话，无需账号或云端。
编报时的 README 依据
<img src=".github/assets/hero.jpg" alt="One Mac's extended screen in a window on another Mac, the Mac's desktop extended onto a tablet, and TwinDesk on a phone" width="100%"> </p> Your devices are good at their own jobs and bad at working together. TwinDesk fixes that. Put a **Mac and a Windows PC** side by side and use them with one mouse and keyboard. Open the PC's apps on the Mac. Turn an old laptop or your **Android tablet** into a second screen. Answer your **phone's calls and texts** from the computer, use its **camera in a video call**, and paste
### 💻 [Computers](#-computer-to-computer) One mouse and keyboard · second screen · the other computer's apps in windows · files · clipboard · **camera** · sound · terminal </td> <td width="33%" valign="top"> ### 📱 [Phones and tablets](#-phone-and-tablet) Calls · texts · photos · files · **camera** · **remote and gamepad** · screen and apps in windows · tablet as a second screen </td> <td width="33%" valign="top"> ### 📡 [Connections](#-connections) Same Wi-Fi · USB cable · **Bluetooth** · **Wi-Fi Direct** and **hotspot** links made on demand · picked for you, moved for you </td> </tr>
[How it works](#how-it-works) · [build and run](#build-and-run) on macOS, Windows, Linux, NixOS and Android · [GitHub Actions](#github-actions) </td> </tr> </table> <a name="download"></a> ## ⬇️ Download <table> <tr> <td width="25%" align="center" valign="top"> **🍎 macOS**<br><sub>13 Ventura or later</sub> <a href="https://github.com/InukaAsith/TwinDesk/releases/download/v2.4.0/TwinDesk-2.4.0-arm64.dmg"><img src="https://img.shields.io/badge/Apple_Silicon-.dmg-000000?style=for-the-badge" alt="macOS, Apple Silicon"></a><br>
<sub>The full app adds notifications, texts and Accessibility; the computer can install it for you over wireless debugging</sub> </td> </tr> </table> <p align="center">🍺 <b>Homebrew</b> on macOS: <code>brew install --cask inukaasith/twindesk/twindesk</code><br> <sub>Update with <code>brew upgrade --cask twindesk</code>: TwinDesk opens again by itself if it was running, so a Mac you drive only through it gets its mouse back. The <a href="https://github.com/InukaAsith/homebrew-twindesk">tap</a> follows every release.</sub></p> <a name="first-launch"></a> <details>
<a href="https://twin-desk.vercel.app"><b>Website</b></a> · <a href="#-new-in-23"><b>What's new</b></a> · <a href="#-get-started"><b>Get started</b></a> · <a href="#how-it-works"><b>How it works</b></a> </p> <p align="center"> <img alt="Version" src="https://img.shields.io/github/v/release/InukaAsith/TwinDesk?label=version&color=4b8bff"> <img alt="macOS 13+" src="https://img.shields.io/badge/macOS-13%2B-black"> <img alt="Windows 10/11" src="https://img.shields.io/badge/Windows-10%20%7C%2011-0078d4">
<img alt="Linux X11 and Wayland" src="https://img.shields.io/badge/Linux-X11%20%7C%20Wayland-f0a30a"> <img alt="Android 10+" src="https://img.shields.io/badge/Android-10%2B-3ddc84"> <img alt="License: GPL-3.0" src="https://img.shields.io/badge/license-GPL--3.0-blue"> <a href="https://github.com/inukaasith/twindesk/releases"><img src="https://img.shields.io/github/downloads/inukaasith/twindesk/total" alt="Downloads"></a> </p> <p align="center">
<summary><b>⚠️ First launch: the apps aren't code-signed</b> <sub>how to allow them on macOS, Windows and Android</sub></summary> TwinDesk is free and has no paid Apple Developer ID or Windows code-signing certificate, so macOS and Windows don't know who made it and ask once before opening it. The source is all here. **🍎 macOS** The easiest way is **Homebrew**, which installs it ready to open: `brew install --cask inukaasith/twindesk/twindesk` From the .dmg instead: 1. Drag TwinDesk into Applications and open it. macOS says it can't verify TwinDesk: click **Done** (not Move to Bin).
- **One notification pool.** Your phone buzzes, and every computer shows it. - **Pair once, join all.** A new device paired with one computer meets the rest by itself. Mesh is in **beta** on Windows, Mac and Linux: turn it on in the **Multi-device** tab, and off again any time with everything as you left it. Connecting just two devices? You won't need it. Still coming to Mesh: the Camera, pictures on the clipboard, and files dragged to more than one device at a time; the phone connects to one computer at a time, and Linux joins over the same Wi-Fi only. --- # 📱 Phone and tablet
- patch-gg/pixelfxJavaScript · 2,641 Stars
特效制作、精灵表导出、AI辅助编程
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
244 pixel art effects, every one a small piece of code. Plus a Claude Code skill to make your own.
编报时的 AI 中文解读
pixelfx 是一个包含 232 个像素艺术特效和动画的库，每个特效都是一小段代码，可直接用于浏览器游戏或导出精灵表到其他引擎，并提供 Claude Code 技能来创建新特效。
编报时的 README 依据
Every effect is a file in `src/fx/<category>/<name>.js`. The [catalog](docs/CATALOG.md) lists them all. ## use one. **In a browser game** (canvas, Phaser, PixiJS, Kaplay, three.js), use the code. No dependencies. ```js import { createFrame, clearFrame, createFx, composite } from "pixelfx"; import * as frostNova from "pixelfx/fx/ice/frost-nova"; const frame = createFrame(320, 180); // your game's pixel resolution const img = new ImageData(320, 180); const fx = createFx(frostNova, { x: 160, y: 140 }); // also: flip, angle, seed, palette, light... function tick(dt) {
fx.update(dt); clearFrame(frame); fx.draw(frame, "back"); // behind characters drawCharacters(); fx.draw(frame, "front"); composite(frame, img.data, backgroundColor); // color, then light added on top ctx.putImageData(img, 0, 0); // scale up with image-rendering: pixelated } ``` Install from GitHub with `npm i github:patch-gg/pixelfx`, or copy `src/core` and the effects you want. **In any engine** (Godot, Unity, GameMaker, Defold, LÖVE, Bevy), use sprite sheets: ```bash node tools/render.mjs ice/frost-nova --sheet
node tools/new.mjs arcane/my-spell # a working starting point (add --loop for a loop) node tools/render.mjs arcane/my-spell # records/arcane/my-spell/contact.png: look at it node tools/check.mjs arcane/my-spell # spills, seams, unseeded randomness, slow frames ``` Then go round: look at the contact sheet, fix the worst thing, render again. `--gif` and `--mp4` show the motion, `--all` shows every frame. [docs/AUTHORING.md](docs/AUTHORING.md) has the contract, the drawing API and the quality bar. `node tools/serve.mjs` runs the gallery locally with your effects in it.
## how it works. - **Two kinds of effect, no images.** Pure effects export `draw(p, t, o)`, the frame at `t` seconds from `t` alone, so loops tile with no seam. Simulated effects export `create(o)` and step real particles, debris and lights at 120 Hz, so sparks bounce and smoke rolls the same at any frame rate. - **Two buffers.** `color` takes pixels: opaque, dithered or truly translucent. `light` takes additive glow, added on top. Hot things look hot because of the light buffer.
- t3dotgg/tstackPython · 2,590 Stars
PR审查、PR合并、PR现代化
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
The agent skills I use every day. Works with Claude Code and Codex.
编报时的 AI 中文解读
这是一套日常使用的代理技能集合，适用于Claude Code和Codex，可帮助处理PR审查、合并、现代化等任务。
编报时的 README 依据
cp -R tstack/skills/* ~/.claude/skills/ # Claude Code cp -R tstack/skills/* ~/.codex/skills/ # Codex ``` ## Requirements - An authenticated [GitHub CLI](https://cli.github.com) (`gh`). - `python3` for `babysit-pr`'s `pr-wait` script. - `council` and `sol-loop` need subagents on other providers' models. T3 Code does this with `delegate_task`. Otherwise you need a harness that can run those models. ## Make them yours These are starting points. Edit them. Put personal settings, like which models sit on your council, in your `CLAUDE.md` or `AGENTS.md`, so you can update the
- Djihads80/EglantineC++ · 2,012 Stars
游戏模拟
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
VERY experimental Nintendo Switch 2 emulator
编报时的 AI 中文解读
Eglantine是一款免费开源的Nintendo Switch 2模拟器，基于Eden、Yuzu和Sudachi开发，用C++编写，支持Windows、Linux、macOS、Android、FreeBSD等平台，旨在运行商业游戏。
编报时的 README 依据
<h4 align="center"><b>Eglantine</b> is a free and opensource (FOSS) Switch 2 emulator, derived from Eden, Yuzu and Sudachi - started by developer Camille LaVey. It's written in C++ with portability in mind, with builds for Windows, Linux, macOS, Android, FreeBSD and more. </h4> <p align="center"> </a> <a href="https://discord.gg/HstXbPch7X"> <img src="https://img.shields.io/discord/1367654015269339267?color=5865F2&label=Eden&logo=discord&logoColor=white" alt="Discord"> </a> <a href="https://stt.gg/qKgFEAbH">
<img src="https://img.shields.io/revolt/invite/qKgFEAbH?color=d61f3a&label=Stoat" alt="Stoat"> </a> </p> <p align="center"> <a href="#compatibility">Compatibility</a> | <a href="#development">Development</a> | <a href="#building">Building</a> | <a href="#download">Download</a> | <a href="#support">Support</a> | <a href="#license">License</a> </p> ## Compatibility The emulator is capable of running most commercial games at full speed, provided you meet the necessary hardware requirements.
We have a user manual! See our [User Handbook](./docs/user/README.md). ## Building See the [General Build Guide](docs/Build.md) For information on provided development tooling, see the [Tools directory](./tools) ## Download You can download the latest releases from [here](https://git.eden-emu.dev/eden-emu/eden/releases). Save us some bandwidth! We have [mirrors available](./docs/user/ThirdParty.md#mirrors) as well. ## Support If you enjoy the project and would like to support us financially, please check out our developers' [donation pages](https://eden-emu.dev/donations)!
- paipiange/nvzhuangGo · 1,876 Stars
服装参考图浏览、AI换装试穿
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
1407 张 AI 女装参考图集与虚拟试穿网站，支持中英检索、OpenAI Images / Gemini / 豆包及自定义 Base URL、API Key、模型。https://github.com/paipiange/nvzhuang
编报时的 AI 中文解读
这是一个收录1407张AI生成女装参考图的网站，支持中英文检索、分类筛选和在线浏览，并可选接入AI换装后端实现虚拟试穿。适合服装设计、电商广告等需要服装参考素材的用户。
编报时的 README 依据
> 🐧 **社区友链**：[LINUX DO (https://linux.do/)](https://linux.do/) ## ⏱️ 10 秒了解本项目 - 🆓 **免费使用（无需部署 / 无需 Key）**：1407 套 AI 服装参考图库浏览、分类筛选、中英双语双向检索、大图全屏查看、基础搭配台与同款跳转，由纯前端静态驱动，在线打开即可使用。 - ⚙️ **自选扩展（需自建后端 / 需自备 Key）**：真实 AI 试衣换装功能依赖 `backend/` 的 GoFrame 服务。需自行配置大模型（如 GPT Image 2、豆包 Seedream 或 Gemini）的 API Key 与 Base URL，调用将消耗对应的第三方模型配额。 - 📜 **完全可商用 · 免费开源**：基于 [MIT License](LICENSE) 全量开源，代码与图集资源完全开放，支持自由商用创作、设计参考及二次开发。详见 [合规审计报告](docs/LINUXDO_RELEASE_AUDIT.md)。 ## 📸 预览 | 中文模式 | 英文模式 | 关键词搜索 | |---|---|---|
### 🚀 后端部署（优先推荐 Docker） 后端推荐使用 **Docker / Docker Compose** 部署，无需在宿主机安装 Go 编译环境，开箱即用。 #### 方式一：Docker Compose（推荐，最省心） ```bash # 1. 复制环境变量模板并填入你的 Key（如国内免翻墙的豆包 ARK_API_KEY，或 OpenAI Key） cp backend/.env.example backend/.env # 2. 一键启动后端 docker compose up -d # 服务启动后监听在 http://localhost:8000 ``` #### 方式二：Docker Run 单命令运行 ```bash cd backend # 构建镜像 docker build -f manifest/docker/Dockerfile -t costume-tryon . # 运行容器（按需注入模型 Key，例如豆包或 OpenAI） docker run -d --name costume-tryon \ -p 8000:8000 \ -e ARK_API_KEY="你的豆包ArkKey" \ -e OPENAI_API_KEY="sk-你的OpenAIKey" \ costume-tryon ```
│ └── img/ # 展示图（webp，弹窗放大用） ├── docs/ # 文档图片与试穿效果演示图 ├── backend/ # AI 换装后端（GoFrame v2，可选） │ ├── docker-compose.yml # 后端 Docker Compose 配置 │ ├── .env.example # 环境变量示例文件 │ ├── restart.bat / build.bat # Windows 本地一键启动/重新编译脚本 │ ├── manifest/config/config.yaml # 模型配置（Key / BaseURL / 模型名） │ ├── internal/logic/tryon/ # 模型适配器（openai / doubao / gemini） │ └── DEPLOY.md # 部署指南 └── .gitignore ``` > 注：原尺寸 PNG（约 1GB+）未纳入仓库，「下载原尺寸」按钮指向原站资源。 ## 🔗 URL 参数
- qybaihe/blaJavaScript · 1,834 Stars
软件防护
Stars 统计更新于 2026/11/2 09:55:22
查看原始资料
作者描述
🛡️ BLA (Blind LLM Agents) — Adversarial defense shield that protects binaries, Electron apps, JS bundles, Python, and WebAssembly against AI reverse-engineering agents like REA. Poison -> Exhaust -> Hallucinate.
编报时的 AI 中文解读
BLA 是一款防御性工具，用于保护开发者自己的软件（如二进制文件、Electron 应用、JS 包、Python 和 WebAssembly）免受 AI 逆向工程代理（REA）的攻击，通过投毒、耗尽和幻觉机制使逆向分析失效，同时保持运行时行为不变。
编报时的 README 依据
[Quick start](#quick-start) · [Empirical results](#empirical-results-vs-real-rea) · [Format coverage](#format-coverage) · [Benchmark](#benchmark-14-targets-internal-estimates) · [FAQ](#faq) ```bash bla harden . # or tell your agent: "Harden my software against REA" ``` </div> > [!IMPORTANT] > BLA is a **defensive** tool for protecting **your own** software. See [DISCLAIMER.md](DISCLAIMER.md). > Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or [npm `bla-agents`](https://www.npmjs.com/package/bla-agents). <div align="center">
| **Hallucinate** | Leave the analyzer with truncated inventories, redaction failures, and ambiguous rabbit holes — while **runtime stays equivalent**. | One sentence covers the common case (auto-discovers `dist/` / `build/` / `.asar` / `.wasm` / `.sqlite` / models / … and runs audit → protect → verify): ```bash npx bla-agents harden . # MCP / agent: bla_harden · "Harden my software against REA" ``` Same pipeline as `bla defend` / `bla_defend` when you already know the path; `harden` is the default one-command UX. ## Quick start ```bash
# Node.js ^22.23.4 or 24.x recommended (REA red-team needs zlib.zstd on 22.23+) npx bla-agents setup # register MCP + skill with your coding agent npx bla-agents harden . # one-command Counter-REA pass on the current project npx bla-agents benchmark # 14-target showcase (local estimates + runtime checks) ``` Ask your agent after setup: ```text Harden my software against REA ``` Or run stages explicitly: `audit` · `protect` · `verify` · `defend` — see [CLI and MCP](docs/cli-and-mcp.md). --- ## Empirical results vs real REA
Recorded under **Node v24.22.1** with optional **Wabt 1.0.42** + **tshark 4.4** when noted. Soft-skips keep Node 22 / CI green when tooling is absent. | Format | Before → after | REA outcome after BLA | Budget / error | | --- | ---: | --- | --- | | **Go buildinfo** | 50 KiB → 1.16 MiB | capacity error | `MAX_BUILD_INFO_BYTES` (1 MiB) — *Go build-info string length exceeds the 1 MiB limit* | | **Erlang BEAM** | 84 B → 262 KiB | capacity error | `max_chunks` — *BEAM form exceeds the 64-chunk budget* |
### Designed-for but **not** empirically verified here These shields exist in-tree and are covered by unit tests / catalog drift checks, but the matching REA path needs **native tooling we did not run** on this box: | Area | Why unverified here | | --- | --- | | Ghidra / native decompilation depth | Requires a Ghidra install + REA bridge | | Frida / LLDB / GDB / Rizin live sessions | Requires those debuggers + device/process attach | | JADX / full Android APK pipeline | APK resigning unsupported; DEX shield is in-tree |
| **Firmware** containers | Skipped — staging/extraction budgets exist; no honest E2E without firmware corpora | | **SPIR-V / CUDA fatbinary / AMD GPU objects** | Skipped — need native GPU object decoders; **PTX text** is shielded in-tree (benchmark only, not a REA GPU-object budget hit) | | EVMole in-process EVM decode | Documented as not pathable in-process; EVM shield still densifies to the 4 MiB input approach | We do **not** claim those paths are empirically bankrupt until `redteam:rea` (or an equivalent harness) records a real before/after against the live analyzer. ---
- haydenschmitty/floorplanJavaScript · 1,805 Stars
代码库可视化、代码导航、健康检查
Stars 统计更新于 2026/11/2 09:35:46
查看原始资料
作者描述
See how your app actually fits together. An interactive, plain-English map of your codebase, drawn by Claude or Codex.
编报时的 AI 中文解读
Floorplan 是一个由 Claude 驱动的代码库可视化工具，能够生成交互式地图，用通俗语言展示应用的各个部分及其连接关系，帮助开发者理解、导航和提问。
编报时的 README 依据
- **For builders:** understand the app you made, in your own words, without reading the code. - **For teams:** onboard someone in minutes, from the big picture down to the exact file. - **For working with Claude:** after every change, see exactly which parts it touched and why. ## Features <table> <tr> <td width="50%" valign="top"> **🗺️ A map at every zoom**<br> Start with the whole app, then double-click into a part, a group of files, a file, and its endpoints. Each view shows what it's used by and what it uses. </td> <td width="50%" valign="top"> **💬 Plain English everywhere**<br>
"Takes card payments at checkout", not "Stripe API wrapper". Technical terms are explained on hover, and one switch shows the developer view. </td> </tr> <tr> <td valign="top"> **🔌 Real connections, across repos**<br> It follows imports, and matches each web request in your frontend to the server endpoint that answers it, even across separate repos. </td> <td valign="top"> **🖥️ Start from a screen**<br> Pick a page you know and see everything behind it: its parts, the requests it makes, the server code, the data, and outside services. </td> </tr> <tr> <td valign="top">
</td> </tr> </table> ## Install You need [Claude Code](https://code.claude.com/docs/en/overview) (the desktop app's Code tab or the CLI) and [Node.js 18+](https://nodejs.org). ```bash claude plugin marketplace add haydenschmitty/floorplan claude plugin install floorplan@floorplan ``` Start a new session and it's ready. Installing the Claude Code CLI is recommended even if you use the desktop app, because it lets the map's Ask bar read your real code. More in [Getting started](docs/getting-started.md). ## Quick start
1. Open your project in Claude Code. For a frontend and backend in separate repos, open the folder that holds both. 2. Ask: > Make me a floorplan of this app. 3. Explore. The map opens on its own, starting with a short tour. When you've done more work, ask: > Update the floorplan. Only what changed is re-read, and what you just worked on lights up under **Recent work**. ### Things to try | Ask Claude | Ask the map | |---|---| | "Make me a floorplan of this app" | "What happens when someone places an order?" | | "Update the floorplan" | "Where do we talk to Stripe?" |
<summary><b>How long does it take, and what does it cost?</b></summary> A first map takes a few minutes for a small app and longer for a big one, because Claude describes every file and endpoint. It uses your Claude plan like any other task. Big projects are split across parallel helpers. Updates re-read only what changed, so they're quick and cheap. </details> <details> <summary><b>Do I need the Claude Code CLI?</b></summary>
- lyramakesmusic/neuron-explorerPython · 1,803 Stars
神经元探索、模型可视化、实时推理
Stars 统计更新于 2026/11/2 09:40:49
查看原始资料
作者描述
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
编报时的 AI 中文解读
这是一个用于探索Qwen3.5-4B模型神经元的交互式工具，提供3D视图、列表、嵌入等多种视角，支持实时推理和Claude集成，适合模型研究者深入分析神经元行为。
编报时的 README 依据
# neuron explorer Steering-first neuron explorer for Qwen3.5-4B. Every MLP neuron in a layer is pushed up and down on a few prompts, every rollout is kept, and the explorer lays the results out over a 3D view of the model.  ## Explorer uv sync uv run python explorer/server.py # http://localhost:8791 Reads sweep runs from `runs/` and per-layer stats from `stats/`, both produced by the sweep below; the caches in `explorer/cache/` ship with the repo. Loads Qwen3.5-4B locally (MPS, CUDA or CPU) for live
steered chat, lens, connections and live inference over your own text. The R and J++ lens rows read the maps of [bubskee/qwen3.5-4b-r-jpp-lenses](https://huggingface.co/bubskee/qwen3.5-4b-r-jpp-lenses) (Apache-2.0): hf download bubskee/qwen3.5-4b-r-jpp-lenses --local-dir explorer/cache/rjpp - **map**: the model's layers in perspective, attention and MLP per layer, one point per neuron. Hover shows a neuron's steer lens and R lens; click opens it; click a layer's label to view its sheet face-on. `writes` places neurons with similar down-projection columns next to each other.
live model (steering, activations, next-token logprobs, ablations), moves the map, list and highlights, and has file tools and a shell in your home folder. Its prompt is `explorer/agent_prompt.md` plus `explorer/agent_methodology.md`, and `explorer/agent_prompt.local.md` when present (untracked, for your own additions); `EXPLORER_CLAUDE_MODEL` picks the model. The server answers only pages and tools on this machine. ## Sweep Runs on a CUDA 13 GPU pod with the steering fork of vLLM ([RhizoNymph/vllm](https://github.com/RhizoNymph/vllm)).
- CopilotKit/OpenDuetTypeScript · 1,794 Stars
主动任务执行、邮件与日程管理、审批与安全控制
查看原始资料
作者描述
An Open Source template for a proactive AI agent that gets your work done and keeps you in the loop.
编报时的 AI 中文解读
OpenDuet 是一个开源、自托管的主动型 AI 智能体，能在网页、移动端、Slack 和 Microsoft Teams 中使用。它可主动提供晨间简报、分类邮件、执行任务，并在发送、支付、签名或删除前请求批准。 它专为个人或团队设计，用户可自行托管并选择自己的模型。
- andyhuo520/digital-human-kouboTypeScript · 1,650 Stars
数字人视频生成、视频剪辑、视频包装
查看原始资料
作者描述
AI digital-human talking-head videos from one photo: Renoise MCP generation + 5 editing styles + 48 creator-grade packaging components. Agent skill for Claude Code / Codex / OpenClaw / WorkBuddy …
编报时的 AI 中文解读
这是一个AI数字人口播视频生成工具，通过一张照片和一句话即可自动生成完整的口播视频，包括脚本、唇形同步、配音、字幕和包装，适用于直播带货、产品推荐、教程讲解等场景。
- oscar-soirey/RavenC · 1,643 Stars
GUI开发、跨语言绑定、可视化界面设计
查看原始资料
作者描述
A lightweight, multithreaded GUI library in C for editors and tools. For C, Rust, C++, C#, Java, Python, Zig, Lua, Go, Odin
编报时的 AI 中文解读
Raven 是一个轻量级、多线程的 C 语言 GUI 库，专为编辑器和工具设计，支持多种语言绑定，并提供可视化 IDE 和 XMLC 编译器，用于快速开发跨平台桌面应用。
- DUNHKpcc/Claude-GuardianPython · 1,627 Stars
账号安全、网络配置、环境搭建
查看原始资料
作者描述
Claude Guardian is an all-in-one Claude account safety and risk prevention skill.Claude防封实战 It intelligently assesses your device, account history, network environment, and payment options, then provides personalized guidance for local, VPS, or virtual machine setups.
编报时的 AI 中文解读
Claude-Guardian 是一个面向 Claude 用户的账号安全与风险预防 Skill，提供订阅准备、账号注册、网络环境配置、本机/VM/VPS 路线选择等指导，帮助用户降低封号风险。
- lexifina/global-lawHTML · 1,594 Stars
法律查询、法律历史追溯、法律文本搜索
查看原始资料
作者描述
Git for Global Law: the law of every country as one git history, each country's record a submodule, every commit one law taking effect
编报时的 AI 中文解读
将全球多国法律整理为一个可搜索的Git历史仓库，每个国家的法律作为子模块，每次提交代表一部法律生效，支持按日期查询法律文本。 提供命令行工具和MCP服务器，便于法律专业人士、研究人员或AI模型查询特定日期的法律条文及其变更历史。
- Pomerado/pomeradoTypeScript · 1,509 Stars
网站集成
查看原始资料
作者描述
YC F26. Turn any website into an API / MCP in minutes. Open source core integration minting harness. MIT.
编报时的 AI 中文解读
Pomerado 是一个将网站转换为 MCP 集成的工具，用户可以用自然语言描述任务，无需录制 HAR 文件即可创建集成，适用于需要让 AI 代理可靠地操作网站的场景。
- windows-GPU-UNLOCK/Windows-GPU-UNLOCKC · 1,482 Stars
GPU解锁
查看原始资料
作者描述
Windows GPU UNLOCK: RTX3090 UEFI port, Korean no-reset Windows boot guide, verified source and benchmark evidence
编报时的 AI 中文解读
实验性UEFI GPU解锁项目，通过自定义EFI引导实现Windows下RTX显卡的直通和性能释放，面向特定硬件配置的高级用户和研究人员。
- TecnaGamer/BeamNG.EarthPython · 1,449 Stars
游戏地图、真实世界驾驶
查看原始资料
作者描述
A 1:1 scale streamed Earth for BeamNG.drive: real elevation and satellite imagery, floating origin, LOD streaming from a companion server
编报时的 AI 中文解读
为BeamNG.drive游戏提供1:1真实地球地图的模组，玩家可在真实海拔和卫星影像上驾驶，支持流式加载、真实天气和多人游戏。
- zhuyansen/awesome-claude-video-skills未标注语言 · 1,426 Stars
视频生成、视频剪辑、技能资源导航
查看原始资料
作者描述
Open-source skills and toolkits that let Claude Code, Codex and other coding agents make video. 180 repos by type, each security-graded. English / 中文.
编报时的 AI 中文解读
这是一个精选的资源汇总列表，收录了能让Claude Code、Codex等编码代理生成视频的开源技能与工具包，共347个仓库，涵盖框架、宣传片、解说、剪辑、短视频、虚拟形象、动画等类型，并提供了安全评级和实测对比。
- herdrdev/herdr-appRust · 1,411 Stars
终端管理、远程开发
查看原始资料
作者描述
herdr desktop app
编报时的 AI 中文解读
Herdr 是一款桌面应用，用于组织终端和编码代理的工作空间，支持多标签页和窗格，并能远程连接其他机器管理终端。适合需要同时处理多个终端会话和代理任务的开发人员。
- Motionsites-AI/claude-skillPython · 1,407 Stars
网页构建、提示词生成
查看原始资料
作者描述
Free Claude skill: make Claude build award-level, cinematic landing pages. By MotionSites.
编报时的 AI 中文解读
这是一个Claude技能，帮助Claude构建电影级落地页，包含示例提示词、媒体素材、技术配方和自动检查修复循环。
- mikufoxxx/Wenku8_FlutterDart · 1,401 Stars
在线阅读、离线阅读、听书
查看原始资料
作者描述
A free Wenku8 and local-book reader for Android, iOS and HarmonyOS, with offline reading, text-to-speech, reading reports, widgets and S3 sync.
编报时的 AI 中文解读
轻小说文库是一款面向Android的第三方阅读客户端，支持在线阅读Wenku8内容及本地书籍导入，提供离线阅读、听书、阅读报告等功能。
- HazyResearch/ncmPython · 1,400 Stars
并行文本生成
查看原始资料
作者描述
NCM: Next Concept Modeling
编报时的 AI 中文解读
Next-Concept Modeling (NCM) 是一个用于并行文本生成的两阶段框架，通过多尺度残差VQ分词器将序列表示为离散码的层次结构，并由Transformer从粗到细预测该层次，实现快速、并行的文本生成。适用于研究并行生成方法、评估模型性能的研究人员。
- cth9191/animateJavaScript · 1,328 Stars
视频生成、动画制作
查看原始资料
作者描述
Procedural animation in any style for Claude Code: short single-file canvas videos with story, look and storyboard check-ins. 7 built-in styles or one matched from your references, your own music (beat-mapped), real product screenshots, every format from one piece.
编报时的 AI 中文解读
这是一个用于Claude Code的提示词技能，通过编写代码生成短片动画视频，支持多种内置风格和自定义风格，适用于制作说明、历史故事等。
- daveebbelaar/bamaSwift · 1,318 Stars
AI辅助编程、应用开发
查看原始资料
作者描述
BAMA: Build Any Mac App. Agent skills to research, clone, design, build, optimize, and test native macOS apps with Swift and SwiftUI.
编报时的 AI 中文解读
BAMA 是一套面向 AI 编程代理的 Skills 集合，帮助代理将 macOS 应用从想法推进到完成，包括克隆现有应用、规划、编码、设计和测试等阶段。适用于使用支持 Skills 的 AI 代理（如 Claude Code、Codex）进行 macOS 原生应用开发的开发者。
- smithjohn78390/tongjiang-labHTML · 1,274 Stars
3D科学课程
查看原始资料
作者描述
Free interactive 3D science lessons for high school · 铜匠实验室 · https://lab.tongjiang.app
编报时的 AI 中文解读
提供适合高中物理、化学、生物、地理学科的免费交互式3D课程，共55节课，每节课为约1MB的独立HTML文件，支持在线或离线使用，界面为中文，遵循中国高中课程标准和PEP 2019教材。
- anonnerdfrenzy/survival-llmPython · 1,243 Stars
离线问答、生存知识查询
查看原始资料
作者描述
A free AI assistant and survival library that runs on your own computer, with no internet needed. Shows the source of every answer.
编报时的 AI 中文解读
这是一个完全离线运行的AI助手，基于内置的生存知识库回答医疗、应急等实用问题，并展示答案来源。适合需要无网络环境下获取可靠生存指导的用户。
- Arjun-Ingole/hopPowerShell · 1,218 Stars
Linux安装、双系统安装
查看原始资料
作者描述
Install Linux next to Windows without a USB stick
编报时的 AI 中文解读
hop 是一个 Windows 应用，用于在已有 Windows 系统旁边安装 Linux，无需 U 盘。用户选择发行版、调整分区大小后点击安装即可，适合希望双系统安装的普通用户。
- Jakeschincariol/arena-skillPython · 1,187 Stars
答案优化、AI辅助编程
查看原始资料
作者描述
When Claude keeps giving you bad answers, make 100 versions of it fight to the death. Same task, 100 different strategies, a bracket, one answer left. Free Claude Code skill.
编报时的 AI 中文解读
这是一个Claude Code技能，通过让100个AI子代理以不同策略竞争同一任务，选出最优答案。适用于对Claude的回答不满意时，通过淘汰赛机制获得更高质量的解决方案。
- nevertoday/xiaoxiaodong-vpsMDX · 1,179 Stars
VPS部署、网络配置
查看原始资料
作者描述
小小东 VPS：3x-ui、Xray、OpenClash、Shadowrocket、CLIProxyAPI 与自动升级部署蓝图
编报时的 AI 中文解读
这是一个帮助用户自行配置海外VPS服务器以访问常用网站和AI工具的方案，包含部署脚本和图文教程，适合需要稳定网络连接的个人用户。
其余 219 个项目见收录列表。