游戏移植、系统优化与AI趣味应用：今日GitHub项目巡礼

收录数按本期首次达标收录的项目统计。

• chuiziyao/maimen-certificationPython · 1,099 Stars

  图像识别认证、套餐推荐、优惠券领取

  Stars 统计更新于 2026/10/25 07:50:39

  查看原始资料

  作者描述

  🍔 麦门认证机 — 上传任意图片，AI 鉴定麦门身份：认证梗图卡 + 真实营养数据套餐推荐 + 一键领券。2026 麦当劳程序员创意开发大赛参赛作品

  编报时的 AI 中文解读

  基于麦当劳MCP接口的趣味AI应用，上传图片生成麦门认证梗图卡，并推荐套餐、领取优惠券、参与抽奖。面向麦门文化爱好者及麦当劳活动参与者。

  编报时的 README 依据

  前往 [open.mcd.cn/mcp](https://open.mcd.cn/mcp) → 手机号登录 → 右上角「控制台」→「激活」→ 复制 Token。 ### 2. 配置 MCP 连接器（WorkBuddy） 左侧边栏【专家·技能·连接器】→【连接器】→ 右上角【自定义连接器】→【配置MCP】： ```json { "mcpServers": { "mcd-mcp": { "type": "streamablehttp", "url": "https://mcp.mcd.cn/mcp-servers/mcd-mcp", "headers": { "Authorization": "Bearer YOUR_MCP_TOKEN" } } } } ``` > ⚠️ 将 `YOUR_MCP_TOKEN` 替换为你的真实 Token。配置文件请勿提交真实凭证。 ### 3. 安装 Skill 把本目录整个文件夹放入 WorkBuddy 技能目录（`~/.workbuddy/skills/`），或在对话中说「用麦门认证机鉴定这张图」。 ### 4. 依赖 - Python 3.x + Pillow（认证卡渲染）

  - Windows 自带微软雅黑 / 黑体字体（其他系统请在 `scripts/generate_card.py` 的 `FONT_DIR` 改为系统中文字体路径） ## 项目结构 ``` ├── SKILL.md # Skill 定义与六步工作流 ├── scripts/ │ └── generate_card.py # PIL 认证卡生成器（750×1080，自适应布局） ├── references/ │ ├── memes.md # 10 档身份体系 + 麦门文学梗语库 │ └── mcd_tools.md # 麦当劳 MCP 工具速查与使用纪律 ├── CONTEST_DECLARATION.md # 参赛声明 ├── MCP_INTEGRATION.md # MCP 集成说明 └── output/ # 生成的认证卡 ``` ## 安全与边界 - 领券、抽奖、下单均为**确认制**：先展示内容，用户明确同意才执行 - 不主动调用下单类工具，支付始终由用户完成 - 证件照、他人隐私照拒绝认证 - 梗语库不含真人、政治、低俗内容 ## 参赛声明

  查看 README 缓存打开 GitHub
• rushiranpise/Shizuku-NextKotlin · 1,036 Stars

  系统API调用、应用权限管理、ADB模块支持

  Stars 统计更新于 2026/10/25 07:40:40

  查看原始资料

  作者描述

  A feature-rich Shizuku fork with improved setup, reliability, UI, and additional start methods.

  编报时的 AI 中文解读

  Shizuku-Next 是 Shizuku 的增强分支，允许应用通过 ADB 或 root 使用系统 API，提供改进的设置、可靠性和界面。适用于需要管理应用权限或使用系统级功能的 Android 用户。

  编报时的 README 依据

  > **Important:** This fork uses a different signing key. Uninstall the official Shizuku or thedjchi's fork before installing Shizuku-Next. A server started by the previous app may remain active until stopped or the device is rebooted. If you want the original app, visit [RikkaApps/Shizuku](https://github.com/RikkaApps/Shizuku). ## Download Get the latest build from [Releases](https://github.com/rushiranpise/Shizuku-Next/releases). ## Features ### Starting and reliability - Start through wireless debugging, USB debugging, root, or Samsung system UID where supported.

  Some features are experimental or device-dependent. Root, Samsung system-UID, and no-Wi-Fi startup methods are not available on every device. Only install modules and scripts you trust. ## Requirements - **Android 7.0 or newer** - **Wireless debugging:** Android 11 or newer; supported Android TVs are also compatible. - **USB/PC start:** A computer with ADB. - **Root start:** A rooted device with working root access. - **System UID start:** Supported Samsung devices only. - **Start on boot:** Supported start methods only.

  查看 README 缓存打开 GitHub
• ZeldaWWHDRecomp/ZeldaWWHDRecompC++ · 1,028 Stars

  游戏运行

  Stars 统计更新于 2026/10/25 07:35:51

  查看原始资料

  作者描述

  Zelda Wind Waker HD Recomp

  编报时的 AI 中文解读

  《塞尔达传说：风之杖HD》的原生移植版，可在macOS、Linux、Windows和Android上运行，无需模拟器。适合拥有正版游戏并希望在现代设备上以更高性能游玩的玩家。

  编报时的 README 依据

  only exist once it is built from your own dump, so the first start prepares the game once, on your computer (about two minutes); every later start launches the game directly. 1. Download the zip for your system from the [Releases](https://github.com/ZeldaWWHDRecomp/ZeldaWWHDRecomp/releases) page and unzip it anywhere (a games folder, an external drive): - **macOS**: Apple Silicon, macOS 14 or newer (Metal renderer; Vulkan through MoltenVK, included, can be picked in Graphics > Renderer)

  and start it from anywhere, Steam Deck included. Nothing is unzipped and the file itself is never written to. 2. Start **Wind Waker HD** (`Wind Waker HD.app`, `Wind Waker HD.exe`, or `wind-waker-hd` / `Wind Waker HD.desktop` on Linux). The first start asks for: - your **disc image** (`.wux` or `.wud`), a **Cemu archive** (`.wua`), or an already **extracted game folder** (with `code`, `content` and `meta`, e.g. from dumpling or Cemu). An extracted folder is used where it is, nothing is copied; a disc image or Cemu archive is extracted into the

  To use it you need your own, legally obtained copy of the game, dumped from your own Wii U disc and console. Everything game-specific (the extracted files, the recompiled code in `build/gen/`, shader caches) is generated locally on your machine from your dump, and must not be redistributed. The `.gitignore` keeps all of it out of the repository. ## Install (releases) Releases are portable: unzip, start **Wind Waker HD**, choose your dump. Releases contain only this project's runtime, tools and setup: **no game files, no game code and no keys**. The game's code can

  release folder (about 1.7 GB); - for a disc image, its **disc key** (a `.key` file with the image's name next to it is used automatically) and the **Wii U common key** (16 bytes, the same on every console; choose a key file or paste the 32 hex digits into the hidden field; a `common.key` next to the image or in the release folder is used automatically). Keys are checked before anything is extracted, never stored, and not part of any log. A Cemu archive or an extracted folder needs no keys.

  offers to bring in a save: a Wind Waker HD `cking.sav` folder (Cemu, Wii U), a GameCube `.gci` (converted to HD), or the saves and settings of an earlier installation or another Wind Waker HD folder (copied, never moved). The USA version (title 00050000-10143500) and the European one (00050000-10143600) are supported, version 0 of either (the disc or eShop release, without the update): before translating, setup checks the game's code (`code/cking.rpx`) against the SHA-256 of each and explains what to use instead when it matches neither (e.g. a game folder with an

  **Development builds and releases include the setup app** (`…-android-arm64.apk`, no game code in it): install it, choose your game, and setup builds the game on the phone. For players: [docs/android.md](docs/android.md) explains what the Android version does, the tested phones and how to play ([Português](docs/android.pt.md), [Español](docs/android.es.md)). You need: - a phone with arm64, Android 13 or newer and Vulkan 1.3; - **a game controller** (Bluetooth or USB) is best: it is the GamePad's buttons and sticks.

  查看 README 缓存打开 GitHub
• JustAResearcher/gpu-unlock-adaC · 719 Stars

  GPU性能解锁

  Stars 统计更新于 2026/10/25 07:50:39

  查看原始资料

  作者描述

  GPU tensor unlock source for RTX 4090 and experimental RTX 4070 Ti SUPER; firmware generated locally

  编报时的 AI 中文解读

  为特定NVIDIA显卡提供Linux内核模块，在NVIDIA驱动加载前解锁SM速度选择，提高张量运算性能。适用于拥有RTX 4090或RTX 4070 Ti SUPER显卡并愿意承担实验风险的用户。 该模块通过自定义固件在驱动前覆盖速度选择，实测RTX 4090 FP16/BF16张量吞吐提升约2倍，但INT8等运算变化不大。

  编报时的 README 依据

  [firmware/README.md](firmware/README.md) for the input and asset layout. Prebuilt firmware packages are not distributed by this repository. ## Enable at boot The package installs the module and firmware into the initramfs so the unlock runs before NVIDIA binds the card. It remains inactive unless the kernel command line opts in: ```text gpu_unlock.targets=0000:03:00.0 gpu_unlock.targets=0000:03:00.0,0000:04:00.0 gpu_unlock.targets=all gpu_unlock.targets=0000:01:00.0 gpu_unlock.ad103_experimental=1 ``` These PCI addresses are examples. Use the full PCI names printed by `lspci -D`.

  `make install` detects the supported distribution family. Its explicit targets are: ```sh make install-arch # makepkg -sif make install-deb # apt-get build-dep, dpkg-buildpackage, apt-get install make install-rpm # dnf builddep, rpmbuild, dnf install ``` `make arch`, `make deb` and `make rpm` build without installing. Packages are built from the committed `HEAD`; Debian and RPM outputs go to `build/deb` and `build/rpm`. Kernel headers and a working DKMS toolchain are required. Builds fetch the pinned NVIDIA `610.43.03` source and binary inputs over HTTPS

  and verify their SHA-256 hashes. `make inputs` prepares the cache; `make firmware` generates assets from cached inputs; `make assets` does both. The default cache is `$HOME/.cache/gpu-unlock` and can be overridden with `CACHE=`, `INPUTS=` and `ASSETS=`. A clean-chroot Debian build needs `GPU_UNLOCK_INPUTS` pointing to the input cache because `dpkg-buildpackage` does not fetch inputs itself. The source archive produced by `make dist`, the Debian source package and the `.nosrc.rpm` contain no NVIDIA archives or generated firmware. See

  needs a card whose firmware state is fresh before NVIDIA loads. A warm reboot is not guaranteed to provide that state: it did on the tested RTX 4070 Ti SUPER, where the pre-firmware WPR2 readback returned to zero and the unlock succeeded again without an AC cut. Other boards may retain WPR2; the module refuses to rerun the firmware sequence in that state. It also skips cards that are already bound, powered down or bus mastering. Loading the module manually after NVIDIA has bound the card cannot unlock it. After validating a cold boot and a one-time warm reboot on the exact card, a

  查看 README 缓存打开 GitHub
• raymondchins/BridgeplayScala · 426 Stars

  Windows游戏运行

  Stars 统计更新于 2026/10/25 07:50:42

  查看原始资料

  作者描述

  Free fork of NotProton: Windows Steam games in Mac Steam, on CodeWeavers' Wine and Apple's Game Porting Toolkit

  编报时的 AI 中文解读

  Bridgeplay 是一个在 Apple Silicon Mac 上通过 Steam 运行 Windows 游戏的工具，它作为 Steam Play 兼容工具集成在原生 macOS Steam 客户端中，让 Windows 游戏像 Mac 游戏一样直接点击 Play 运行。

  编报时的 README 依据

  3. **Follow the Setup Assistant**: Rosetta, graphics, Steam sign-in and a couple of standard macOS prompts. 4. **Press Play** on any Windows game in Steam. The first time you use a controller, allow Steam's Input Monitoring request. Each game's Bridgeplay settings (graphics backend, Metal HUD, DLSS, frame generation, controllers) are in Steam: right-click the game, **Properties… › Bridgeplay**. Updates install themselves: Bridgeplay, its patch in Steam and the engine. **Refresh** on the Status page checks on demand.

  | Mac | Apple Silicon, M1 or later. Intel Macs are not supported. | | macOS | 26 or later | | Steam | A Steam account. Windows games install from Steam as usual. | | Download | About 570 MB. Installing Rosetta 2 and Steam, and signing in to Steam, need internet. | ## How it works Bridgeplay runs inside the native macOS Steam client and registers itself as a Steam Play compatibility tool. Under the hood: - **Engine:** CodeWeavers' open-source WineCX 11.17 with MSync, running under Rosetta 2.

  Also running on the same setup: Street Fighter 6, Mortal Kombat 1, The Last of Us Part I, Death Stranding Director's Cut, Horizon Zero Dawn, Split Fiction, It Takes Two, A Way Out, Batman: Arkham Knight, Sifu, Ghostrunner, Control Resonant, Ravenswatch, Megabonk, PowerWash Simulator 2, The Witcher 3. _{Games with kernel-level anti-cheat do not run.} ### Compatibility list

  查看 README 缓存打开 GitHub
• stencil-hq/dquicktsC · 400 Stars

  持久化执行、TypeScript运行

  Stars 统计更新于 2026/10/25 07:50:42

  查看原始资料

  作者描述

  QuickJS with durable execution: suspend running JS, snapshot the whole heap, resume it in another process. Native TypeScript, a REPL that survives restarts, Rust bindings.

  编报时的 AI 中文解读

  这是一个基于QuickJS的JavaScript引擎分支，增加了持久化执行和原生TypeScript支持，适用于需要跨进程恢复运行状态的场景。

  编报时的 README 依据

  git clone https://github.com/stencil-hq/dquickts cd dquickts make # builds qjs, qjsc and libquickjs.a ./qjs hello.ts ./qjs --drepl session.snap # a REPL whose state survives restarts make test-snapshot # builds and runs the suspension and snapshot test hosts ``` ## Contents - [What this fork adds](#what-this-fork-adds) - [Durable REPL](#durable-repl) - [Snapshots](#snapshots) - [What a snapshot is](#what-a-snapshot-is) - [When you can take one](#when-you-can-take-one) - [Example: hibernating in the middle of a loop](#example-hibernating-in-the-middle-of-a-loop)

  - The `std` and `os` modules are not available. Their files, timers and handlers are host state that a snapshot cannot save. - `\load file` evaluates a file inside the session. - Ctrl-C interrupts a running command. The session is not affected. - A session file only loads into the `qjs` build that wrote it. - Tab completion lists properties of the global object. Like the regular REPL, it does not list top-level `let`, `const` and `class` names, but it does complete their members (`c.in` → `c.inc()`). ## Snapshots ### What a snapshot is

  a restore: the interrupt handler, module loader, memory and stack limits, promise rejection tracker, context opaque pointers, and locale callbacks. Modules imported after the restore go through the new runtime's loader. - **Buffers.** `ArrayBuffer` data is always restored into memory the engine owns. Host buffers are copied, and their free functions are not kept. ### Limitations - `JS_WriteSnapshot()` throws a `TypeError` for: - contexts with the Intl or Temporal intrinsics, - `SharedArrayBuffer`s whose memory may be shared with another thread,

  - host-class objects with opaque data when there is no `save_opaque()` hook, - C closures (`JS_NewCClosure()`), whose opaque pointer is host state, - any call made while JS code is running on the C stack. - A runtime can have only one suspended execution at a time. - If a suspended frame cannot be resumed (stack overflow, or the interrupt handler interrupts it), it and its callees are terminated without running `catch`/`finally` blocks or `using` disposals. The error is thrown at the frame's call site.

  查看 README 缓存打开 GitHub
• Alterssss/dsh-bili-whaleJavaScript · 357 Stars

  视频浏览、日记生成、视频分享

  Stars 统计更新于 2026/10/25 07:50:42

  查看原始资料

  作者描述

  给 DSH 装一条有日子的鱼：自己按作息刷 B 站、真看懂视频、写两本日记，把值得的 @ 你。

  编报时的 AI 中文解读

  这是一个为 DSH 宿主开发的插件，让一个 AI 角色按设定作息自动浏览 B 站视频、理解内容、撰写两本日记，并选择性地通过评论或私信分享给你。它强调克制和低风险，适合希望有个性化 AI 陪伴、自动筛选视频内容的 B 站用户。

  编报时的 README 依据

  | **私信专线**（`share.via = dm` + `notify.type = dm`）| **两个不同的、可登录的实名账号**，且**双方手动互关** | ❌ 做不到 —— 见下 | > ⚠️ 国内一个身份证通常只能实名一个 B 站号 ✗ > 所以「第二个号」对很多人是**过不了实名**的硬墙 ✗ > 只有一个号时：**分享方式选「评论区 @」，推送选桌面通知 / Webhook / 自定义命令** ✓ > 单号模式**不能**做：私信专线、以完全独立的第二个身份出现 ✗ ## 装 **前提**：Node 18+；**ffmpeg**（看画面要用，见下）。 ```bash # 1) 装插件（本地目录 / npm 都行） dsh plugin --profile desktop add "link:C:\path\to\dsh-bili-whale" # 2) 重启一次 DSH（bundle 层要重启才组装） # 3) 打开「设置 → 养鱼」（不是 /dsh-bili/ 那个页面 —— 那个被 DSH 的信任栅栏挡着 ✗） ``` **设置页里三件事**：

  设置保存在 `~/.dsh/bili-whale.config.json`（**不含凭据**）；凭据单独在 `~/.bili-ai/credentials.json`，**永远不回显**。 --- ## 🚦 启动测试（装完**立刻**跑这一条） ```bash node <插件目录>/tools/bili.mjs selftest ``` （Windows 路径换成 `%USERPROFILE%\.dsh\plugins\dsh-bili-whale\tools\bili.mjs`） 它会把该确认的事列成一张表：配置 / 凭据 / 登录态 / 数据目录 / 依赖体检， 缺哪项就直说缺哪项，**最后两行是绝大多数人卡住的地方**： ``` ────────── 接下来要做的两件事（很多人卡在这）────────── ① 【找到她的会话】 插件启动时会自动开一个会话，位置在： 左侧「工作区」→ 展开「未分组」文件夹 → 里面那个 bili-whale ⚠️ 不点开「未分组」是看不见它的 —— 这一步最容易漏 ✗ ② 【给她完全权限】 打开那个会话 → 权限设置 → 选「完全访问 / danger-full-access」 否则她起来后跑不了命令、发不了评论、写不了日记 ✗

  node --test tests/planner.test.mjs tests/integration.test.mjs tests/updates.test.mjs tests/supplement.test.mjs ``` 57 个用例，覆盖： - 计划器：跨午夜、绝对时间戳、跳过不重抽、重启后计划仍在、按真实截止排序、旧状态迁移 - 分享策略：开关 / 评分 / 理由 / 排名，**评论与私信两条通道都测** - 集成：自定义 `DSH_HOME`、人设注入、日记推送开关、审批失败后草稿恢复、 Node 运行时缺失时报错、调度器的午夜行为与清理 > 测试用**临时目录 + 虚构凭据 + 模拟网络与宿主接口** —— 不登录真实账号，也不发真实消息 ✓ > 它验证的是逻辑，不是真实 DSH 版本兼容性和 B 站接口可用性 ✗ 那两样只能装上去跑。 --- ## 许可 MIT（见 `LICENSE`）。第三方与免责见 [`NOTICE.md`](NOTICE.md)。 本项目是**个人自用工具的公开版**，与 B 站、DeepSeek 官方无隶属关系；用你自己的账号做事，后果自负。 <!-- 2026-10-21 issue#8 fix --> ## 安装前提（2026-10-21 补，Issue #8） **这个插件必须挂在带网页界面的组件层上运行。**

  # dsh-bili-whale 0.4.6 > 给 DSH 装一条**有日子的鱼**：它自己按作息刷 B 站、真看懂视频、写两本日记，把值得的 **@ 你**。 > > 不是"定时任务 + 关键词搜索"，是**生物钟 + 偏好 + 克制**：每天只醒一两次，筛五条、深读两三条、**最多 @ 你三次**。 她该长成什么，见 [VISION.md](VISION.md)；每一版的改动见 [CHANGELOG.md](CHANGELOG.md)。 ## ⚠️ 默认值是按「刚创的小号」定的（先看这个） 新装用户的默认配置是**保守**的。原因很直接：**长时间高频轮询是小号被风控的头号原因** ✗ | 项 | 默认值 | 为什么 | |---|---|---| | **评论监听** | **关** ✗ | 开了就是每 60 秒一次请求，一天 1440 次 ✗ 小号扛不住 | | **私信监听** | **开** ✓ 但 **60 秒一轮** | 私信是她唯一的"叫得应"通道，但间隔不能短 | | **全天监听** | **关** ✗ | 只在作息窗口（默认 20:00–24:00）内听 | | **同一个人只回一条** | **开** ✓ | 30 天内不重复骚扰同一个人 | | **请求间隔** | **1500 ms** | 连击是最容易被判机器人的特征 |

  `~/.dsh/AGENTS.md` 会被 DSH **在所有会话里热加载** —— 所以它对**每一个** agent 生效， 不只对她 ✗ 想只改她就用上面的 `persona.md` ✓ 这里适合放**通用红线**（不说假话、不冒充你、不伤害第三方）和跨端的协作约定。 ### 三个字段也能改 设置页里那三个（角色名 / 怎么称呼你 / 怎么自称）会拼成一行身份提示： ``` ［身份］你是「…」，称用户「…」，自称「…」。 ``` 只是称呼提醒，放不下详细人设 ✓ 详细内容写 `persona.md` ✓ > 旧版本的 `persona.example.md` 已移除：那是**某个具体的人**的人设，不该当模板发出来 ✗ --- ## 账号前提（先看这个，再决定用哪条路） 这个插件是**两个号的模型**：「她」一个号，「你（主人）」一个号。同一个号做不到"她跟你说话"。 | 通道 | 需要什么 | 只有一个号时怎么办 | | --- | --- | --- | | 看视频 / 写日记 / 学偏好 | 只要一个号（她的 cookie）| ✅ 完全可用 | | 分享（`share.via = comment`）| 同上，不需要第二个号 | ✅ 完全可用（在评论区 @ 你）| | 推送（桌面 / Webhook / 自定义命令）| 同上 | ✅ 完全可用 |

  | `B站私信` | **不发公开评论**，把分享内容 + **视频链接**私信给你 | 不想吵到别人 ✓（就是评论区那 110 条里喊"艾特泛滥"的人要的 ✓）| > 两种方式**共用同一本分享额度**：≤3 条/天 · 同一视频 7 天一次 · 评分 ≥13/20 ✓ > 判重按 **BV 号**算（不是按"发给谁"算）✓ **两条通道走同一套校验**（v0.4.0 起）—— 关掉分享总开关，私信也发不出去： | 校验 | 说明 | |---|---| | `share.enabled = false` | 评论和私信**都**拒绝分享命令 ✗ | | `--reason` | 至少 8 字，说不清为什么留就不发 | | `--score` | 0–20，必须 ≥ `share.minScore`（默认 13）| | `--rank 1` | 只允许第一名 | > 以前私信那条**绕过**了这些校验（关掉分享还能发）✗ v0.4.0 修了 ✓ > 普通聊天私信（不分享视频）不受这些限制 —— 它不需要评分 ✓ --- ## 专属通道（私信）—— 只有一条路，写死在代码里 想把东西**私密地**送到你手里、又不想在评论区吵到别人？用 `dm`（**一天同一个人不限条数** ✓）： ```bash

  查看 README 缓存打开 GitHub
• actionow-ai/vvvPython · 325 Stars

  视频剪辑、AI辅助编程、视频模板复用

  Stars 统计更新于 2026/10/25 07:50:42

  查看原始资料

  作者描述

  Build a video template once, reuse it across coding agents. Editable Remotion templates and an Agent Skill for Claude Code, Codex and 25+ agents.

  编报时的 AI 中文解读

  提供可复用的视频模板和Agent Skill，让用户通过对话让编码代理自动完成视频制作，包括选模板、放照片、写文案、渲染和检查。

  编报时的 README 依据

  Picking the template, placing photos, writing copy, rendering, checking and delivering all happen in the conversation. You watch the sample and say what to change. > Use the vvv skill to make a 9:16 wedding short from the photos in ./photos. We are Lin Yizhou and Xu Zhixia, married on October 18. Show me a checked sample first. **Add it to your agent**, any one of: | Agent | Install | |---|---| | **Claude Code** | `/plugin install vvv --marketplace actionow-ai/vvv` (the plugin includes the templates) |

  | **Codex, Cursor, Gemini CLI, Copilot and 20+ more** | `npx skills add actionow-ai/vvv`, then clone this repository and set `VVV_KIT` ([details](docs/agent-compatibility.md)) | | **WorkBuddy** | Upload the skill ZIP from [Releases](https://github.com/actionow-ai/vvv/releases) under 技能 → 添加技能 ([details](docs/workbuddy.md)) | | **Any MCP client** | `claude mcp add --transport http vvv https://mcp.vibevision.video/mcp` to browse templates and guides with no install | Rendering runs on your machine, which needs Python 3.9+, Node.js 26 and npm; the agent installs everything else.

  <td width="33%" valign="top"><b>Status</b><br>0.1, an early release. Verified end to end on macOS ARM64 with Claude Code and Codex; WorkBuddy and other platforms are less tested. <a href="docs/verification.md">Verification and limits</a></td> <td width="33%" valign="top"><b>License</b><br>Code, docs, templates and demo media are <a href="LICENSE">MIT</a>. Fonts, preview soundtracks and third-party code keep their own terms: <a href="MEDIA_LICENSE.md">media</a> · <a href="THIRD_PARTY_NOTICES.md">third-party notices</a></td>

  Best with an image-capable agent. With Claude Code and Codex both installed, Claude Code can use Codex's image generation through the bundled bridge. Video and voice are not generated. [How it works](docs/capabilities.md#recreate-a-reference-video) ## Also included - **Render checks:** `vvv check` measures format, audio, still stretches and beat alignment, then lists what only a viewer can judge. - **Handoff:** pack a template, effect or finished project as one [`.vvv` file](docs/format/vvv-1.0.md) with its code, media licenses and render evidence, so any agent can continue it.

  查看 README 缓存打开 GitHub
• LNPANDY/NX2AppleC · 315 Stars

  游戏转制、RomFS归档

  Stars 统计更新于 2026/10/25 07:50:42

  查看原始资料

  作者描述

  Play your own Switch games on iPhone/iPad — GUI fork of nxnative: RomFS archiving, 48 games, auto-sync

  编报时的 AI 中文解读

  这是一个在iPhone/iPad上运行自己拥有的Switch游戏的工具，提供macOS图形界面和命令行两种方式，将游戏转制成iOS可安装的IPA包，并支持大游戏RomFS归档以解决安装问题。

  编报时的 README 依据

  # NX2Apple **Play your own Nintendo Switch games on an iPhone or iPad.** Point a macOS app at your own game dumps and it runs the whole pipeline: unpack, rewrite the game's code for Apple silicon, translate its shaders to Metal, and pack a slim IPA plus a sharded RomFS archive that a phone can actually install — then tells you the three steps to get it onto the device. <img src="frontend/screenshots/main-window-en.png" width="620" alt="NS Port Builder"> The project is based on **[nxnative](https://github.com/seedds/nxnative)** — the NSO rewriter, the

  - **The IPA does not install.** A 16.6 GB package with 372,660 entries fails on iOS with `PackageExtractionFailed (Could not extract archive)`. The install daemon's extractor gives up — one directory holds 75,941 files and the ZIP64 central directory starts at 17.8 GB. - **Copying the tree to the phone takes hours.** The Files app and Finder do one round trip per file. 300,000 small files also produce roughly 1 GB of filesystem block overhead. Packing the tree into ~100 shard files solves both: the package stays small, and the game files move to the phone as one folder of large files.

  Horizon host, the NVN-on-Metal layer and the per-game profiles are upstream work, and this fork follows upstream automatically. What it adds: - **A GUI app for everyone.** [NS Port Builder](frontend/README.md) walks a first-time user through the whole build: four inputs picked separately (or one folder for all), device dependencies detected and installed with one click, progress from the build's own stages, and an install guide with Reveal-in-Finder buttons at the end. English by default; Chinese on Chinese devices.

  DLC, `prod.keys`), a dependency panel that detects and installs what is missing, the script's real stage list with live percentages, and the archive shards produced at the end. <details> <summary>界面截图（点开）</summary> <img src="frontend/screenshots/main-window.png" width="620" alt="NS Port Builder"> </details> ```bash cd frontend && swift run # or double-click frontend/NS Port Builder.app ``` The interface is bilingual: it follows the device (Chinese for any Chinese variant, **English for everything else**) and the `中文` / `EN` button in the title bar switches at any time.

  [README](https://github.com/seedds/nxnative) to install Xcode, Homebrew tools and the Python packages, then pack an IPA. ```bash git clone https://github.com/<you>/NX2Apple.git cd NX2Apple brew install xcodegen zstd spirv-cross brew install --cask dotnet-sdk python3 -m pip install lz4 capstone fonttools backports.zstd ``` ### 2. What you need on disk Three things, none of which ship with the repository: | What | Where it comes from | Used by | |---|---|---| | **`prod.keys`** | dumped from **your own console** (see below) | the extractor, to decrypt your game |

  | **The game's NSPs** | dumped from your own cartridge or eShop purchases | the extractor | | **Xcode + iOS SDK** | the App Store | the app build | `prod.keys` is the console's key set — a text file of `name = hex value` lines (~230 entries: `master_key_xx`, `titlekek_xx`, `header_key`, `key_area_key_application_xx`, …). Every Switch homebrew tool reads this format. It is **not** tied to a specific game: one `prod.keys` decrypts any of your games. The only requirement is that the file contains keys for the key generation

  **Sideloadly** does not. See [dm/ios/ENTITLEMENTS.md](dm/ios/ENTITLEMENTS.md) for the steps and how to tell from `Documents/log.txt` (resident + headroom: ~3,375 MB = not in effect, 6–8 GB = it is). ## Limitations - **Read-only.** The archive is never written; saves and temp data go to their own mounts as before. - **Case-sensitive lookup.** Names are stored as the source tree had them, so the host's case-insensitive alias machinery (for APFS) is bypassed for archive entries. - **Symlinks are followed at pack time.** No supported game's RomFS has any.

  - **The index is held in memory.** 28 MB for Tears of the Kingdom (17.9 MB of it names). Shards are not. - **No dedup or compression.** Shards are the source bytes verbatim, so `verify` is a plain comparison and a repack is deterministic. ## Credits All of the emulation work is other people's. This repository adds the archive subsystem on top: - **[nxnative](https://github.com/seedds/nxnative)** (MIT) — the native-execution toolchain this repository is built from: the NSO rewriter, the Horizon host layer, NVN-on-Metal, the per-game profiles and the iOS shell.

  查看 README 缓存打开 GitHub
• ivarvong/lanternTypeScript · 201 Stars

  动画制作、代码讲解

  Stars 统计更新于 2026/10/25 07:50:45

  查看原始资料

  作者描述

  Animated, steppable code explainers. Zero-dependency TypeScript, SVG output, browser and Cloudflare Workers.

  编报时的 AI 中文解读

  lantern 是一个零依赖的 TypeScript 库，用于创建动画化、可逐步讲解的代码变更说明，输出 SVG 格式，可在浏览器、Cloudflare Workers 或 Node 中运行。适合开发者制作教学演示或交互式文档。

  编报时的 README 依据

  st.cache = true; // ... } } ``` Register it in `scenes/index.ts`, then: ``` npm run see -- my-scene after # the frame as text, plus the step's state diff npm run shot -- my-scene # contact sheet: the end of every step, as a PNG npm run verify:motion -- my-scene # teleports, pops, cuts, flicker ``` Diagrams can be text: `flowchart(mermaidSource)` parses a Mermaid flowchart, lays it out and returns shapes you can animate ([docs/FLOWCHART.md](docs/FLOWCHART.md)). Scenes can also be data: `sceneFromDoc(doc)` plays a JSON document of nodes, edges,

  | one frame as SVG | `renderAt(Scene, t, new SnapshotSink())`, or Worker `GET /frame.svg?scene=&step=` | | one frame as text | `renderText(scene)`, `npm run see`, or Worker `GET /text?scene=&step=` | | state at every step | `traceScene(Scene)`, or Worker `GET /trace.json?scene=` | | every frame | `npm run render -- <scene>` writes `out/frames/<scene>/*.svg` | Deploy: `npm run deploy` (builds `out/site/`, then `wrangler deploy`). ## Commands | | | |---|---| | `npm run dev` | source tree as native ES modules, live reload |

  # lantern Animated, steppable explainers for code changes. A zero-dependency JS port of manim's core model that renders to SVG: a live DOM in the browser, strings in Cloudflare Workers or Node. Every step is addressable (`?scene=&step=`), and every scene has a text form and a structured trace for agents. ## Quickstart ``` git clone https://github.com/ivarvong/lantern && cd lantern # not on npm yet; zero runtime dependencies ``` Then: The repo is TypeScript throughout. Node 24 runs the tools and tests as `.ts` directly

  (type stripping, so only erasable syntax); `tsc` compiles the library to `dist/`. ``` npm install npm run dev # http://localhost:5173/?scene=pyex-limits, reloads on save, keeps your step npm test # ~8 s; FULL=1 (npm run test:full) checks every frame npm run check # tsc over every project, emits dist/ ``` A scene is a class with an async `construct()`. Steps are markers; `this.state` is what the explanation is about, and it is what traces record: ```ts import { Scene, Rectangle, Text, Labeled, Arrow, FadeIn, Create, RIGHT, TEAL, GOLD } from '../src/index.ts';

  | `MANIM_DIR=… npm test` | also diffs rate functions and curve subdivision against manim's own code | | `npm run see`, `shot`, `render` | look at a scene: text, contact sheet, SVG frames | | `npm run trace`, `profile`, `seek`, `ab`, `hitch` | performance ([docs/PERF.md](docs/PERF.md)) | | `npm run fonts` | re-subset the embedded fonts (needs `pip install fonttools brotli`) | ## Out of scope, for now - Matching morphs (`TransformMatchingShapes`): glyphs pair by index, not by shape. - 3D, boolean path ops, LaTeX. - Rigid moves as SVG `transform` attributes: every move rewrites points.

  - Collaboration: the trace and doc formats are diffable, but nothing syncs them yet.

  查看 README 缓存打开 GitHub
• songjianyang/xss-audit未标注语言 · 197 Stars

  XSS审计

  Stars 统计更新于 2026/10/25 07:50:45

  查看原始资料

  作者描述

  Static XSS audit skill with 14 layers of checks for React, TypeScript, and JavaScript projects.

  编报时的 AI 中文解读

  这是一个用于对React、TypeScript和JavaScript项目进行静态XSS漏洞审计的Agent技能，提供14层检查，覆盖DOM注入、事件处理器、URL导航、原型污染、CSS注入、SVG/MathML、Markdown/sanitizer绕过、框架渲染等多个方面，并生成审计报告。适用于安全审计人员或开发者在代码审查中排查XSS风险。

  编报时的 README 依据

  ## Installation Copy this directory into your agent's skills directory: ```text <skills-directory>/xss-audit/SKILL.md ``` ## Usage ```text /xss-audit /xss-audit src/ ``` Or use `$` if that is your agent's skill syntax: ```text $xss-audit $xss-audit src/ ``` ## What it checks The 14 review layers cover: 1. DOM and HTML injection: `innerHTML`, `outerHTML`, `insertAdjacentHTML`, `document.write`, `DOMParser`, `setHTMLUnsafe`, shadow DOM, and React `dangerouslySetInnerHTML`.

  See the [fictional sample report](examples/sample-report.md). It demonstrates the format without claiming to be the result of a real audit. ## Limitations Static analysis can produce false positives and false negatives, especially for dynamic runtime behavior, build-time transforms, excluded code, and framework-specific execution timing. Review exclusions and verify findings before relying on them. No coverage is guaranteed. Generated reports may contain sensitive paths, source code, internal domains, or credentials. Review and sanitize them before sharing. ## Disclaimer

  查看 README 缓存打开 GitHub
• xMazaki/unbloat-android-tvJava · 179 Stars

  系统精简、主屏替换

  Stars 统计更新于 2026/10/25 07:50:45

  查看原始资料

  作者描述

  Slim down any Android TV (ads, telemetry) + minimalist Apple TV-style home screen. Reversible, one double-click. 🇫🇷 Allégez n'importe quelle TV Android (pubs, télémétrie) + écran d'accueil minimaliste façon Apple TV. Réversible, en un double-clic.

  编报时的 AI 中文解读

  这是一个用于精简Android TV系统（去除广告、遥测和不必要服务）并安装极简Apple TV风格主屏幕的工具套件，支持Windows、macOS和Linux，操作可逆。

  编报时的 README 依据

  | RAM disponible | 451 Mo | 720 Mo | | Écran d'accueil | 187 Mo (Google TV) | 25 Mo (Zen Home) | | Services désactivés (pubs, télémétrie, inutiles) | 0 | 32 | Ces chiffres viennent d'une seule TV. Le gain sera différent sur la vôtre. ## Compatibilité | Appareil | Allègement | Zen Home | |---|---|---| | TV **Android TV / Google TV** : TCL, Sony, Philips, Xiaomi, Hisense (modèles Android), Sharp, Toshiba… | ✅ | ✅ | | Boîtiers : Nvidia Shield, Chromecast avec Google TV, Xiaomi Mi Box, Google TV Streamer | ✅ | ✅ |

  | **Amazon Fire TV / Fire Stick** (Fire OS) | ✅ | ❌ Utilisez un outil dédié ([voir ci-dessous](#fire-tv--fire-stick--changer-laccueil)) | | Fire TV sous **Vega OS** (Fire TV Stick 4K Select et modèles sortis depuis fin 2025) | ❌ Pas Android | ❌ | | Samsung (Tizen), LG (webOS), Hisense VIDAA, Roku, Apple TV | ❌ Pas Android | ❌ | **Testé en conditions réelles :** une TCL sous Android 9. Les autres appareils de ce tableau devraient fonctionner, car ils utilisent les mêmes commandes Android, mais ils n'ont pas tous été testés. ### Fire TV / Fire Stick : changer l'accueil

  7. [Personnaliser Zen Home](#personnaliser-zen-home) 8. [Sécurité et retour en arrière](#sécurité-et-retour-en-arrière) 9. [Dépannage](#dépannage) 10. [Contenu du dossier](#contenu-du-dossier) 11. [Responsabilité](#responsabilité) 12. [Crédits et licence](#crédits-et-licence) --- ## En bref 1. Activez le mode développeur et le débogage sur la TV ([voir comment](#préparer-la-tv-5-minutes)). 2. Double-cliquez sur **`LANCER.bat`** (Windows) ou **`LANCER.command`** (macOS), ou lancez **`./LANCER.sh`** (Linux), puis tapez l'adresse IP de la TV. 3. Acceptez « Autoriser le débogage » sur la TV.

  > Dans votre box, réservez cette adresse IP pour la TV (« bail DHCP statique »). Elle ne changera plus. ## Utilisation 1. Double-cliquez sur **`LANCER.bat`** (Windows) ou **`LANCER.command`** (macOS). Sur Linux, ouvrez un terminal dans le dossier du kit et tapez `./LANCER.sh`. 2. Choisissez une option du menu. La première fois, le kit demande l'adresse IP de la TV. 3. **Regardez la TV** : la fenêtre « Autoriser le débogage USB ? » s'affiche. Cochez **« Toujours autoriser depuis cet ordinateur »**, puis **OK**.

  **Le kit ne propose que ce qui existe sur votre TV.** Une Sony ne verra que les lignes Google et Sony, par exemple. Si un paquet est absent ou si la TV refuse de le désactiver (certains sont verrouillés, surtout sur Fire TV), il est simplement ignoré et le kit continue. **Il vous faut :** un ordinateur sous **Windows 10 ou 11**, **macOS** ou **Linux**, connecté au **même réseau** que la TV. Sur Linux, `curl` (ou `wget`) et `unzip` doivent être installés. ADB, l'outil officiel de Google pour parler à Android, est téléchargé automatiquement au premier lancement (environ 7 Mo).

  2. Descendez jusqu'à **Build** (ou **Version de l'OS Android TV**) et **appuyez 7 fois** dessus avec le bouton OK de la télécommande. 3. Le message « Vous êtes maintenant développeur » s'affiche. ### 2. Activer le débogage 1. Ouvrez **Paramètres → Système → Options pour les développeurs**. 2. Activez **Débogage USB**. S'ils existent, activez aussi **Débogage ADB réseau** ou **Débogage sans fil**. ### 3. Noter l'adresse IP de la TV Dans **Paramètres → Réseau et Internet → (votre Wi-Fi)**, ou **À propos → État**, notez l'adresse du type `192.168.1.50`. > [!TIP]

  > **Projet personnel partagé tel quel, sans aucune garantie.** Je ne suis pas responsable des problèmes que son utilisation pourrait causer à votre TV, à vos appareils ou à vos données. Chaque TV est différente : lisez la section [Avant d'utiliser le kit sur votre TV](#avant-dutiliser-le-kit-sur-votre-tv). --- ## Sommaire 1. [En bref](#en-bref) 2. [Ce que fait le kit](#ce-que-fait-le-kit) 3. [Compatibilité](#compatibilité) 4. [Avant d'utiliser le kit sur votre TV](#avant-dutiliser-le-kit-sur-votre-tv) 5. [Préparer la TV (5 minutes)](#préparer-la-tv-5-minutes) 6. [Utilisation](#utilisation)

  ## Avant d'utiliser le kit sur votre TV Chaque fabricant installe ses propres applis, avec des noms différents selon les marques, les modèles et les versions d'Android. Le kit a été conçu pour être prudent, mais il ne peut pas connaître toutes les TV du marché. **Vous devrez peut-être adapter le kit à votre TV.** Par exemple : - ajouter ou retirer des lignes dans `scripts/listes/pubs_telemetrie.txt` et `scripts/listes/optionnel.txt` ; - ajouter l'appli « Entrées TV » de votre marque dans le code de Zen Home, si elle n'est pas reconnue ;

  Amazon protège son écran d'accueil : sans root, impossible de le désactiver ou de le remplacer comme le fait le kit sur les autres TV. Utilisez le kit pour l'**allègement** (options 1, 2 et 6), puis l'un de ces outils dédiés pour l'accueil. Ils redirigent la touche Accueil vers le launcher de votre choix grâce à un service d'accessibilité, sans root, sur Fire OS 7 et 8 : - **[Projectivy Launcher](https://play.google.com/store/apps/details?id=com.spocky.projengmenu)** : un launcher complet, avec l'option « Remplacer le launcher actuel » dans ses réglages.

  查看 README 缓存打开 GitHub
• Charonyuu/whatsmyagentdoingJavaScript · 167 Stars

  代理活动监控、多代理管理

  查看原始资料

  作者描述

  See what every Claude Code and Codex agent on your machine is doing — live, in a little office.

  编报时的 AI 中文解读

  这是一个Claude Code和Codex的Skill，安装后输入/myoffice即可在浏览器中实时查看本机所有Claude Code、Codex和Gemini CLI代理的活动状态，以2D或3D办公室的趣味形式呈现。

  查看 README 缓存打开 GitHub
• jouaaaaaa/mcd-self-butlerPython · 161 Stars

  点单推荐、健康分析、团餐配餐

  查看原始资料

  作者描述

  麦当劳自助管家 McButler — 基于麦当劳 MCP 的点单参谋：每日推荐 / 套餐健康体检 / 门店全成本比价 / 团餐智能配餐 / 麦险分信用评级 / 麦门小剧场。麦当劳 1024 程序员节参赛作品。

  编报时的 AI 中文解读

  基于麦当劳官方MCP的点单参谋工具，提供每日推荐、健康体检、门店比价、团餐配餐、麦险分信用评级和麦门小剧场等功能，面向普通消费者和开发者。

  查看 README 缓存打开 GitHub
• rddmwmddr/android-ad-cleanerPython · 141 Stars

  手机去广告

  查看原始资料

  作者描述

  关掉国产安卓手机系统里的广告、热搜、资讯推荐和营销推送。不 root，可还原，支持 AI 智能体一键执行。

  编报时的 AI 中文解读

  这是一个帮助用户关闭国产安卓手机系统中广告、热搜、资讯推荐和营销推送的工具，提供桌面客户端、AI智能体、聊天AI和手动清单等多种使用方式，无需root，可还原。

  查看 README 缓存打开 GitHub
• xkx121029/mcd-vibe-commitPython · 138 Stars

  人格测试、commit生成、娱乐彩蛋

  查看原始资料

  作者描述

  麦门人格 Commit 生成器：测出你的麦门人格，生成专属 git commit message（麦当劳 MCP 大赛参赛作品）

  编报时的 AI 中文解读

  这是一个娱乐向的 Agent Skill，通过3题测试判定用户的麦门人格，并生成对应的 git commit message。适合开发者、麦当劳爱好者以及想为团队增添趣味的用户。

  查看 README 缓存打开 GitHub
• ravensclawx/HBLauncherJavaScript · 123 Stars

  自制软件启动器、GPU应用SDK

  查看原始资料

  作者描述

  SmileBASIC homebrew launcher and documented GPU app SDK, with Pico setup and a Hello World example.

  查看 README 缓存打开 GitHub
• CH-LaTiao/lumina-guidePython · 104 Stars

  场景理解、通行提示

  查看原始资料

  作者描述

  LUMINA Guide / 灵犀: Jetson + Orbbec RGB-D wearable navigation assistant prototype

  编报时的 AI 中文解读

  面向视障人群的可穿戴RGB-D场景理解与通行提示原型，基于Jetson和Orbbec相机，提供本地感知、语音和振动反馈，当前为公开演示版。

  查看 README 缓存打开 GitHub
• hiyfoeorihly/ram-purgePython · 99 Stars

  内存清理、性能优化

  查看原始资料

  作者描述

  Memory cleaner and RAM booster for Windows: purge standby list, trim working sets, free cached RAM, and auto-clean on threshold.

  编报时的 AI 中文解读

  Memory Cleaner: RAM Booster & PC Optimizer 是一款面向 Windows 10/11 的免费内存清理和优化工具，提供工作集修剪、备用列表清除、修改页刷新等功能，并支持自动调度和静默模式。

  查看 README 缓存打开 GitHub
• big-tooth/mcd-recommend-meal-by-emotion未标注语言 · 92 Stars

  心情荐餐、优惠券匹配、在线下单

  查看原始资料

  编报时的 AI 中文解读

  这是一个基于麦当劳MCP和腾讯WorkBuddy Skill构建的智能荐餐助手，根据用户心情、预算和时段推荐麦当劳餐品，并支持优惠券匹配和下单。适用于使用WorkBuddy的麦当劳会员用户。

  查看 README 缓存打开 GitHub
• PandaSI-Labs/littleaiboxJavaScript · 88 Stars

  聊天辅助、文档阅读、AI辅助编程

  查看原始资料

  编报时的 AI 中文解读

  Little Box 是一个注重隐私的对话工作区，支持聊天、写作、翻译、编程帮助和文档阅读，可在浏览器、PWA 和 Android 应用中使用，用户自带模型密钥，数据默认保存在本地。

  查看 README 缓存打开 GitHub
• lzp76995-labb/mcd-fairbiteJavaScript · 76 Stars

  点餐规划、预算管理

  查看原始资料

  编报时的 AI 中文解读

  为2-6位同事或朋友提供多人点餐协商工具，在共同预算和不同口味下生成每人一份、预算可核查的麦当劳点餐草案。适合需要快速协调多人点餐的场合。

  查看 README 缓存打开 GitHub
• Addition2000/mcd-diet-guardPython · 73 Stars

  忌口点餐、多人点餐

  查看原始资料

  作者描述

  帮有忌口的朋友在麦当劳安心点餐。 不吃猪肉、吃素、海鲜过敏、乳糖不耐、不吃辣、不吃酸黄瓜……说一句话，它就查遍整家门店的实时菜单，告诉你哪些放心点、哪些“去培根”就能吃、套餐里该换成什么，算好价再帮你下单。

  编报时的 AI 中文解读

  麦麦忌口管家是一款帮助有饮食禁忌的用户在麦当劳安全点餐的工具。它通过MCP接口获取实时菜单，根据忌口对餐品进行分层推荐，支持自动特调与套餐换选，并生成下单参数。适合有忌口、过敏或口味偏好的个人及其代点餐者。

  查看 README 缓存打开 GitHub
• brainstormity/System-One-Computer-UsePython · 68 Stars

  电脑自动化

  查看原始资料

  编报时的 AI 中文解读

  这是一个通过截图控制电脑的原型演示工具，利用AI模型模拟点击、输入等操作，适用于探索自动化体验，但仅限测试环境使用。

  查看 README 缓存打开 GitHub
• SkyZeroZx/corner-shopTypeScript · 64 Stars

  安全漏洞演示

  查看原始资料

  作者描述

  Angular 22 SSR + NestJS demo: the trust boundaries of Angular HTTP (DevFest Callao 2026)

  编报时的 AI 中文解读

  这是一个用于演示Angular HTTP信任边界安全漏洞的示例应用，包含Angular SSR、NestJS API和攻击者服务器，通过unsafe和safe两种模式展示并验证安全问题，适用于安全研究者和Angular开发者学习。

  查看 README 缓存打开 GitHub
• kiyoakii/xinliqiu-sdkPython · 57 Stars

  录音取回、音频解密

  查看原始资料

  作者描述

  Unofficial SDK for the OPPO AI 心力球: pair from any computer, sync and decrypt your own recordings, optional ciphertext-only storage, MCP server for agents.

  编报时的 AI 中文解读

  非官方SDK，用于从任何计算机配对OPPO AI心力球设备，同步并解密自己的录音，可选密文存储，并提供MCP服务器供agent使用。 提供Python库和CLI两种接口，用于读取设备状态、取回录音数据、解密音频，并支持通过Cloudflare进行密文存储。

  查看 README 缓存打开 GitHub

查看本期全部收录项目