This repository is a curated collection of Claude skills—markdown instructions and supporting scripts that enable Claude (or any MCP-compatible agent) to perform a wide range of tasks, from document manipulation (DOCX, PPTX, PDF, XLSX) and GIF animation to web testing, MCP server building, and automation of hundreds of third-party SaaS services via Composio's Rube MCP. The repo is primarily a skill library with a standardized template for each skill, but it also contains several well-crafted Python libraries (e.g., easing functions, particle systems, OOXML validation frameworks) that are genuinely reusable. The architecture is modular: each skill lives in its own directory with a SKILL.md and optional scripts. The heavy reliance on external services (Composio, Rube MCP) and the massive duplication of templated skill documentation introduce maintenance risks, but the core utility scripts demonstrate solid engineering craftsmanship.
lx-music-desktop is a cross-platform Electron-based music player (Vue 3 + TypeScript) that aggregates multiple Chinese streaming services (NetEase, Tencent, Kugou, Kuwo, Migu, Baidu, Xiami). It provides music search, playback, playlist management, lyrics (including desktop lyric windows and karaoke-style font rendering), downloads with quality tiers, audio effects (EQ, convolution reverb, pitch shifting), cross-device synchronization over a custom WebSocket protocol, and an HTTP Open API for third-party control. The architecture spans three processes: main (Electron window/tray/IPC, SQLite persistence via better-sqlite3), renderer (Vue app with custom reactive stores and a plugin-based audio engine), and renderer-lyric (dedicated always-on-top lyric window). All built-in music sources were removed in v2.6.0 following a Tencent legal complaint; the app now relies entirely on user-supplied Music API scripts, introducing a sandboxed script runtime with security warnings. The repository is actively maintained (last commit 2026-09-19) but the project is in a transition phase—the maintainer has announced a successor project ('Any Listen'), and the codebase shows signs of legacy accumulation: extensive commented-out code, duplicated components (e.g., MusicSortModal), hardcoded API endpoints/signatures that break frequently, and a custom store implementation diverging from Vuex/Pinia. Despite this, the code is well-organized around provider SDK adapters, composable logic hooks, and a clear main/renderer IPC boundary, with several genuinely reusable patterns.
This repository is a curated resource catalog for WeChat Mini Program development. It does not build a deployable application in the traditional sense; instead, it operates as a structured data project that generates a searchable static site and API from a YAML data file. The core architecture is a set of TypeScript scripts that read `data/resources.yaml`, validate the data integrity, check external links for liveness, and generate `README.md`, `public/api/resources.json`, and the search interface in `public/index.html`. The data is categorized into official docs, tools, plugins, components, backend SDKs, and demos, each with metadata like language, difficulty, and topics. The repository's primary value lies in its data quality tooling. The validation script (`validate-data.ts`) enforces strict schema rules, checks for duplicate URLs, validates metadata, and provides precise YAML line numbers for errors. The link checker (`check-links.ts`) uses concurrent HTTP requests with retries and graceful handling of servers that block HEAD requests. The import script (`import-readme.ts`) can reconstruct the YAML from the generated README, enabling a bidirectional workflow. The GitHub Actions CI pipeline runs validation, generation checks, and link checks on pull requests, posting automated comments when checks fail. The codebase is well-structured with clean separation of concerns: `catalog.ts` handles all data types and YAML/JSON I/O, `generate.ts` handles README and API generation, and individual scripts handle specific tasks. The TypeScript configuration uses strict mode and modern ES2022 features. Dependencies are minimal (only `yaml` at runtime) and the project requires Node.js >= 20. The last commit was in June 2026, indicating the project is in a stable but not actively developed state.
V2Ray-core is a Go-based network proxy and tunneling platform designed to help users build private networks and bypass censorship. It functions as a modular proxy framework supporting multiple protocols (VMess, VLess, Trojan, SOCKS, HTTP, MTProto, Shadowsocks) and transport layers (TCP, mKCP, WebSocket, HTTP/2, QUIC, Domain Socket, XTLS). The architecture is highly modular: the core runtime manages features (dispatcher, router, DNS, stats, policy, log) as pluggable components, while proxy implementations and transports are registered via a global config-creator registry. Entry points are the main binary (v2ray) and a control utility (v2ctl) for config conversion, signing, and diagnostics. Data flows from inbound handlers through a dispatcher that routes connections by destination, protocol sniffing, and routing rules to outbound handlers, which then dial remote servers over the configured transport. External dependencies include gorilla/websocket, quic-go, miekg/dns, go-proxyprotocol, VSign for binary signing, and xtls/go for extended TLS. The codebase demonstrates strong engineering in its layered abstraction: common/ provides reusable, low-level primitives (buffered I/O with MultiBuffer, retry logic, crypto streams, session context, platform path handling), while features/ defines clean interfaces for routing, DNS, stats, and policy that the app/ layer implements. The transport layer is particularly well-designed, with each protocol isolated behind a common Dial/Listen interface and supporting pluggable packet headers for traffic obfuscation (e.g., faking WeChat video, SRTP, UTP, WireGuard, DTLS). The string matcher (strmatcher) uses a trie for domain matching and supports full, substring, keyword, and regex patterns with index-based grouping. The custom buf package provides a lightweight buffer pool and vectored I/O reader (readv) for performance. The error handling uses a structured errors package with path objects for source location. However, the repository shows signs of stagnation. The README explicitly redirects to v2fly/v2ray-core, indicating this is the legacy v2ray organization's repo. The last commit was in April 2026, but the go.mod targets Go 1.15 (released 2020), and many dependencies are pinned to pre-1.0 versions (quic-go v0.18.1, grpc v1.32.0, protobuf v1.25.0). The CI workflows have conditions like `if: github.repository != 'v2ray/v2ray-core'` suggesting they are disabled for this fork. The codebase contains extensive generated code (errors.generated.go, config.pb.go) and mock files. The XTLS integration references an external fork (xtls/go) rather than the standard crypto/tls. The custom Chacha20 implementation uses code generation for the core block function. While the architecture is sound, the dependency versions and Go version target are significantly behind current ecosystem standards.
This repository is a reference architecture and plugin collection for deploying AI agents in financial services using Anthropic's Claude platform. It defines 10+ named agents (e.g., Pitch Agent, Earnings Reviewer, KYC Screener) and 100+ skills that guide Claude to produce institutional-grade financial documents: Excel valuation models (DCF, LBO, comps), PowerPoint pitch decks, Word research reports, and HTML earnings previews. The architecture follows a dual-delivery model: agents can be installed as Claude Cowork plugins via marketplace registration, or deployed headlessly via the Managed Agents API (agent.yaml manifests in managed-agent-cookbooks). Skills are authored in vertical plugins organized by domain (investment banking, equity research, private equity, wealth management) and synced into agent bundles. The repository also includes a Microsoft 365 add-in installer (claude-for-msft-365-install) for deploying Claude into Excel, Word, PowerPoint, and Outlook. Data flows rely heavily on MCP servers (LSEG, S&P Global, Daloopa, FactSet) and SEC EDGAR for financial data, with strict citation and formula-over-hardcode conventions. The repo is primarily prompt engineering and configuration rather than application runtime code, but it includes several reusable Python scripts for skill scaffolding, document validation, and cross-reference linting.
OpenAI Gym is a foundational Python library providing a standard API for Reinforcement Learning (RL) environments. It defines the core `Env` interface (`reset`, `step`, `render`) and includes a registry of environments spanning classic control (CartPole), physics simulations (MuJoCo, Box2D), and grid worlds. The architecture relies heavily on a composition-based wrapper pattern (`gym.wrappers`) to modify environment behavior without altering core logic, and includes a vectorization module (`gym.vector`) for parallel execution using shared memory. While the codebase demonstrates strong architectural craftsmanship in its abstraction layers and space definitions, it is in a maintenance phase with development migrated to Gymnasium. It relies on external compiled libraries (MuJoCo, Box2D, Pygame) which introduce platform-specific complexity and dependency drift.
This repository is a massive, multi-language collection of algorithmic problem solutions primarily from LeetCode, 剑指Offer (Sword to Offer), and Cracking the Coding Interview. It serves as a reference and study resource for coding interview preparation. The core content is organized by problem number and includes solutions in over 10 languages (Python, Java, C++, Go, TypeScript, Rust, etc.). A secondary component is a web application (Dockerfile, package.json) that likely powers the site at leetcode.doocs.org, providing a browsable interface for the solutions. The repository is actively maintained with frequent commits and a large contributor base, but the codebase is dominated by boilerplate solution files with high duplication and no automated testing. The architecture is a flat collection of independent solution files rather than a cohesive software application.
PythonRobotics is a large collection of robotics algorithms implemented in Python, designed primarily for educational purposes and as a companion to a textbook. It covers a wide range of robotics domains including localization (EKF, particle filter, histogram filter), mapping (grid maps, NDT, clustering), SLAM (EKF SLAM, FastSLAM, GraphSLAM), path planning (A*, Dijkstra, RRT variants, PRM, DWA, Frenet, etc.), path tracking (Stanley, LQR, MPC, pure pursuit), arm navigation (inverse kinematics, RRT for arms), aerial navigation (drone trajectory, rocket landing), bipedal walking, inverted pendulum control, and mission planning (state machines, behavior trees). Each algorithm is implemented as a standalone script with a main() function that runs a simulation and visualizes results using matplotlib. The code is not packaged as a library and is not intended for production use; it is a learning resource with clear, readable implementations that reference academic papers. The repository includes extensive documentation in Sphinx RST format, CI/CD pipelines, and a test suite that runs each demo without animation.
Koodo Reader is a cross-platform ebook reader application supporting a wide range of formats (EPUB, PDF, MOBI, AZW3, TXT, FB2, CBR/CBZ, MD, DOCX, HTML, etc.). It is built as an Electron desktop app with a React/TypeScript frontend and a separate Go HTTP server for OPDS catalog serving and KOReader sync. The application provides a full-featured reading experience including bookmarks, notes, highlights, text-to-speech (with multiple voice engines), translation and dictionary lookup (via built-in and custom plugins), OCR for scanned PDFs, and cloud sync/backup across many providers (WebDAV, S3, OneDrive, Google Drive, Dropbox, etc.). It also includes a library management system with shelves, tags, search, and import from local files, folders, URLs, and OPDS catalogs. The architecture is a monolith with a clear separation between the Electron main process (Node.js) and the renderer (React), plus an optional Go server for network-facing features. The codebase is large and feature-rich but shows signs of technical debt: widespread use of deprecated React lifecycle methods, heavy reliance on opaque minified libraries (kookit-extra), frequent `as any` type casts, and mixed Chinese/English comments. Despite these issues, the application is actively maintained and offers a robust set of features for ebook enthusiasts.
MusicFree is a plugin-driven, cross-platform mobile music player built with React Native and Expo. In production it provides audio playback, playlist management, lyric display and fetching, music searching, downloading, local music scanning, theme customization, and backup/restore. The app is not a data pipeline or research project; its only persistent data is user config, playlists, cache metadata, and music history. The architecture is a React Native application with a custom plugin runtime: external CommonJS JS plugins define music-source capabilities such as search, getMediaSource, getLyric, and getMusicSheetInfo, and the app dynamically loads those plugins through a sandboxed require. Core services include a track player wrapper, plugin manager, music sheet manager, downloader, lyric manager, theme manager, and i18n system, all coordinated through Jotai atoms, event emitters, and singleton stores. Entry points are index.js (app registration and TrackPlayer initialization), src/entry/index.tsx (main app shell), and src/entry/bootstrap/bootstrap.ts (startup initialization). External dependencies include react-native-track-player, MMKV, AsyncStorage, react-native-fs, expo modules, WebDAV, react-native-reanimated, and react-native-gesture-handler.
cool-retro-term is a desktop terminal emulator that simulates the visual appearance of vintage cathode-ray tube (CRT) monitors. It renders a real terminal session (via QMLTermWidget) with a chain of GPU shader effects including scanline patterns, bloom, screen curvature, burn-in, chromatic aberration, and frame reflections. The application is built with Qt 6 (QML and C++) and runs on Linux and macOS. It provides a rich settings interface for customizing all visual parameters, and ships with 14 built-in retro profiles (e.g., Amber, Green, Commodore 64, Apple II) and a curated collection of bitmap fonts to match authentic terminal looks. Architecturally, the C++ layer handles font management (FontManager), font list models, file I/O, and the main entry point. The QML layer composes the terminal window, tabs, settings panels, and the shader pipeline. Shaders are pre-compiled into .qsb files for 48+ combinations of effect flags, enabling smooth runtime switching without compilation latency. Persistent settings are stored in SQLite via a thin QML wrapper (Storage.qml). The application uses two Git submodules: qmltermwidget for the actual terminal emulation and KDSingleApplication for single-instance enforcement. Key entry points include app/main.cpp (C++ startup), app/qml/main.qml (QML root), and the shader fragment files in app/shaders/. External dependencies are Qt 6, QMLTermWidget, and KDSingleApplication. The build system is qmake (with a planned but not yet executed migration to CMake). Packaging is provided for AppImage, DMG, Debian, RPM, and Snap.
PandaWiki is a production-ready, self-hosted AI knowledge base and wiki platform. The actual codebase implements document management with versioned releases, RAG-based Q&A and search, AI-assisted document summarization and editing, multi-source content ingestion (URL, RSS, Sitemap, Notion, Feishu, DingTalk, Confluence, EPUB, Wiki.js, Yuque, Siyuan, etc.), a customizable public landing page builder, analytics, comments, contribution review, and multi-channel bot integrations for WeChat, WeCom, DingTalk, Feishu/Lark, and Discord. It is not a data research repository, but it contains substantial document-ingestion and analytics pipelines that feed the RAG system. The architecture is a full-stack monolith-style application split across a Go backend (Echo + GORM + PostgreSQL + Redis + MinIO/S3 + NATS JetStream), a React admin panel (Vite/MUI/Redux), a Next.js public site, and a pnpm workspace of shared UI/icon/theme packages. The Go backend follows a layered design with handlers, use cases, repositories, middleware, and dependency injection via generated Wire code. A standalone MQ consumer process handles async RAG document updates, crawler exports, and cron-driven statistics aggregation. The public app is served separately from the admin panel, with an embeddable widget and an H5 chat interface. In production, documents are imported through crawler tasks or direct upload, converted to HTML/Markdown, chunked, and upserted into a configurable vector/RAG service. User chat requests rewrite the query, retrieve ranked chunks, and stream LLM responses via SSE with citations. Statistics are rolled up hourly into PostgreSQL. Authentication supports local accounts, JWT sessions, GitHub OAuth, LDAP, CAS, and enterprise SSO. Feature availability is gated by commercial license editions, with a private pro submodule adding enterprise functionality.
Rubick is a cross-platform desktop application built with Electron and Vue 3 that serves as an extensible productivity toolbox. Its core functionality includes a global shortcut-triggered launcher (Alt/Option+R) for quick access to applications, files, and folders, a plugin system that treats plugins as installable npm packages, and WebDAV-based cross-device data synchronization via PouchDB. The architecture features a main Electron process managing multiple renderer windows (main search interface, plugin market, detached plugin windows, and tutorial guide) with communication through IPC. Key dependencies include Vue 3, Ant Design UI library, PouchDB for local storage with WebDAV sync, and native modules for global hotkeys and file icons. The plugin system allows users to extend functionality through a built-in market, with plugins stored locally and loaded into isolated BrowserView instances for security.
XianyuAutoAgent is an AI-powered customer service bot for the Chinese second-hand marketplace platform Xianyu (闲鱼). It connects via WebSocket to Xianyu's messaging service, receives user messages, and generates intelligent replies using a Large Language Model (LLM) with multi-expert routing. The system features intent classification (price negotiation, technical support, general query), dynamic pricing negotiation with gradual concessions and adaptive LLM temperature, context-aware conversation management via SQLite, and a manual takeover mode. It handles authentication via cookies, token refresh, heartbeat, and captcha-triggered cookie re-entry. The bot operates as a long-running async process, suitable for deployment via Docker.
Material Files is a sophisticated, open-source Android file manager built with Kotlin and Material Design principles. Its core architecture is a custom implementation of the Java NIO2 file system API (`java.nio.file.spi.FileSystemProvider`), providing a unified abstraction over multiple backends: the local Linux filesystem (accessed via JNI syscalls), Android's proprietary Documents/Content providers, and remote network protocols including SMB, SFTP, FTP, and WebDAV. The UI is constructed from Android Fragments, RecyclerView adapters, and custom Material Design widgets, orchestrated by a ViewModel/LiveData pattern for reactive state management. The application goes far beyond basic file browsing. It features a comprehensive background file job service for copy, move, delete, and archive operations with progress tracking and conflict resolution. It includes a built-in FTP server, an image viewer with EXIF support, a text editor, and deep file property inspection (APK metadata, audio/video tags, checksums, POSIX permissions, and SELinux context). Root access is achieved through a separate process communicated via AIDL (Android Interface Definition Language), demonstrating a sophisticated IPC architecture for privilege escalation. The codebase is actively maintained by a single developer, with recent commits and a clear version history. However, it suffers from high technical drift due to its heavy reliance on Android's internal, undocumented, and frequently changing APIs (e.g., `android.system`, `android.os.storage`). The extensive `compat` package is a testament to the constant battle against platform evolution. This makes the project fragile and high-risk for long-term maintenance without dedicated effort. Despite its overall complexity, the repository contains several genuinely reusable, high-craftsmanship components. Standouts include the `ByteString`/`ByteStringPath` classes for handling non-UTF-8 filesystem paths, a generic `ListDiffer` algorithm, a robust `Stateful`/`DataState` state machine pattern, and clean implementations of throttling/debouncing. These components are self-contained and could be extracted for use in other Kotlin or Java projects. The overall quality score is 7.5, reflecting strong engineering and feature depth, balanced against significant platform-specific technical debt.
This repository is a well-established Python library (`SpeechRecognition`) that provides a unified interface for multiple speech recognition engines, including Google Speech, CMU Sphinx (PocketSphinx), Vosk, OpenAI Whisper (both local and API), Groq, Cohere, and several others. The architecture follows a facade pattern via a central `Recognizer` class, which handles audio capture (microphone, files) and delegates recognition to pluggable backends – each implemented as a standalone function in the `recognizers/` subpackage and attached to the `Recognizer` instance at import time. The library bundles FLAC encoder binaries for audio conversion, offers a CLI tool for model downloads, and includes comprehensive test suites with audio fixtures in multiple formats. While the code is well-structured and features clean abstractions (e.g., `AudioData` for format conversion, `OpenAICompatibleRecognizer` base class), it exhibits some technical drift: bundled FLAC binaries are outdated (v1.3.2 from 2014), it depends on `audioop` (deprecated in Python 3.13+), and several backend APIs are at risk of deprecation. The library is production-grade (v3.17.0) and widely used, but its reliance on hardcoded API keys and legacy workarounds for old Python versions suggests maintenance is ongoing but with moderate lag. Overall, it is a solid, reusable codebase with several components that could be extracted independently.
Mage is an open-source data pipeline orchestration platform that combines a visual notebook-style editor with a production-grade scheduling and execution engine. The system enables users to build, run, and monitor data pipelines composed of Python, SQL, R, and dbt blocks. Architecturally, it is a full-stack web application with a Python backend (Tornado-based server, SQLAlchemy ORM, REST API) and a Next.js/React frontend with a custom component library. Key entry points include the CLI (`mage` command), the web UI served by `mage_ai/server/server.py`, and the API layer at `mage_ai/api/` which exposes CRUD resources for pipelines, blocks, runs, schedules, users, and integrations. Data flows center around pipeline execution: blocks are connected in a DAG, each block can be a data loader, transformer, or exporter. Execution is handled by a configurable executor (local, K8s, EMR, PySpark) and scheduled via a built-in scheduler or external triggers. The platform integrates with dozens of data sources and destinations through a Singer-based integration framework, supports streaming pipelines, and includes advanced features like RBAC, version control, AI-assisted pipeline generation, and a command center for global search. It depends heavily on external services such as cloud storage (S3, GCS, Azure), databases (PostgreSQL, BigQuery, Snowflake, etc.), and authentication providers (OAuth, LDAP, OIDC).
yewtube is a terminal-based YouTube player and downloader, forked from mps-youtube. It provides a full-featured CLI for searching, streaming, downloading, and managing YouTube content without requiring a YouTube API key. The application uses yt-dlp and youtube-search-python as backends for video extraction and search. It supports multiple media players (mpv, mplayer, VLC), MPRIS integration for desktop media control, Last.fm scrobbling, playlist management (local and YouTube), and various download/transcoding options. The architecture is a monolithic Python application with a command-loop pattern, global state management via a dedicated module, and a plugin-like player abstraction. While the codebase shows its age with some tightly coupled modules and global variables, it contains well-designed components such as a config system with validation, an abstract player interface, and a command registration decorator. The project is actively maintained with recent commits and releases.
Awesome Home Assistant is a community-curated index (an awesome list) of third-party resources for the Home Assistant home automation platform. The repository contains a single massive README.md that serves as the canonical list, organized into sections for custom integrations, dashboard cards, themes, apps, DIY projects, community resources, and more. It is not a software application itself but a well-maintained documentation and curation project with a set of Python maintenance scripts and GitHub Actions workflows that validate entries, check for broken links, refresh star counts, and generate a static website via MkDocs. The repository is designed for contributors to propose additions via pull requests; automated checks enforce entry quality criteria (license, age, activity).
MediaElement is a mature, client-side JavaScript library that provides a consistent, cross-browser HTML5 video and audio player experience. It wraps native `<video>` and `<audio>` elements with a unified API and UI, while supporting a plugin-based renderer system for fallback and streaming formats including YouTube, Vimeo, HLS, DASH, Facebook, SoundCloud, Twitch, and Dailymotion. The library is distributed as two bundles: a core `mediaelement.js` (API only) and `mediaelement-and-player.js` (with full UI controls). It includes internationalization (i18n) for over 30 languages, subtitle/caption track support, and a range of player features (play/pause, progress bar, volume, fullscreen, time display) that can be customized via a feature injection pattern. Architecturally, the codebase is organized into core modules (`mejs.js`, `mediaelement.js`, `renderer.js`, `i18n.js`), renderer implementations (each self-contained and registered via a common interface), player features (each added as a method on the player prototype), and utility functions (DOM helpers, time formatting, media type detection, polyfills). The build system uses Grunt with Browserify and Babel 6, targeting older browsers (IE10+, iOS 7+, Android 3+). The library is designed to be framework-agnostic but includes optional jQuery/Zepto integration. While the library is functionally complete and widely used, it shows signs of technical drift: the build toolchain is outdated (Grunt, Browserify 13, Babel 6), some dependencies are stale, and the codebase relies on global namespace pollution (`window.mejs`). The player class (`player.js`) is monolithic and handles many responsibilities. Despite these issues, the core abstractions—especially the renderer registry and i18n engine—are well-designed and could be extracted for reuse in other projects.