Agent Skills is Google's repository of installable skills for Google products and technologies, including Google Cloud. The skills provide Markdown-based procedures for tasks such as onboarding and authenticating to Google Cloud, deploying and managing AI agents, working with GKE, databases, networking, security, monitoring, analytics, advertising APIs, Firebase, Android, Dart, Flutter, and Google Maps Platform. Skills can be selected and installed with `npx skills add google/skills`; the repository also bundles plugins and MCP servers for agent harnesses including Claude Code, Codex, and Antigravity CLI. The repository accepts bug reports and contributions, and its skills are distributed under the Apache 2.0 license.
Angular is an open-source web development platform for building mobile and desktop web applications with TypeScript, JavaScript, and other languages. It provides components and templates, forms, APIs, server-side rendering, lazy loading, animations, and Angular Elements, along with the Angular CLI for creating workspaces, running local development servers, and managing projects. The project is hosted in the angular/angular repository and includes documentation, upgrade guidance, contribution guidelines, and an ecosystem that includes Angular Material.
Claude-Mem is an open-source persistent-memory plugin and service for coding agents. It captures agent activity through lifecycle hooks, stores sessions, observations, and summaries in SQLite, and uses hybrid full-text and Chroma vector search to retrieve relevant context across sessions. Its MCP search workflow uses progressive disclosure: the agent first searches a compact index, then reviews a timeline, and finally fetches full observations for selected result IDs. A local worker service, managed by Bun, provides the HTTP API, search endpoints, and web viewer; the project also supports integrations with Claude Code and other listed agent environments, configurable context injection, private-content exclusion tags, and optional cloud synchronization. The repository states that it is distributed under the Apache License 2.0 and requires Node.js 20 or later, with Bun, uv, and SQLite used by the runtime. It can be installed through its npx installer or Claude Code's plugin marketplace.
Co-op Anything is an open-source MCP server and playbook for directing AI coding agents to add cooperative multiplayer to single-player PC games. The MCP server searches a directory of more than 230 co-op mods, detects a game's engine from its installation folder, provides research and reverse-engineering guidance, and validates contributions. An agent either revives an existing mod by auditing and fixing its code or builds co-op from scratch by hooking into the game, synchronizing players, enemies, worlds, story, level changes, and saves. Builds are tested with bots driving real game clients, then packaged for installation and Steam-relay play without port forwarding. The repository includes agent skills, engine-specific guidance, networking and split-screen playbooks, automated QA procedures, game-specific projects, and contribution rules. It is licensed under the MIT License.
DeepTutor is an open-source, self-hostable AI tutoring platform developed by HKUDS for lifelong personalized tutoring. It provides a web application and command-line interface for interactive learning, problem solving, quiz generation, deep research, visualization, and mastery practice, with persistent memory and learning state shared across knowledge bases, books, notebooks, tutor personas, and other activities. The platform supports multi-agent problem solving, retrieval-augmented learning with source-traced and page-level citations, math animations, interactive visualizations, and tutor bots built from personal study materials. Its knowledge sources include personal documents, EPUBs and books with annotations, GitHub repositories, web search, and connected libraries; documented retrieval and ingestion options include GraphRAG, PageIndex, LightRAG, linked knowledge bases, Obsidian, and configurable parsing and vector backends. DeepTutor also includes courses, research workflows, an ecosystem of MCP services and tool or capability plugins, and integrations with connected coding agents and other partners. It can run locally or through Docker.
djs is a small JavaScript and TypeScript engine written in no_std Rust and built mainly for WebAssembly. It runs codemode scripts, including model-written programs that call tools, process their results, and return data; the project describes it as a QuickJS replacement for pi's codemode. The virtual machine stores all state in one byte array addressed by 32-bit offsets, allowing a new VM to be created by copying a prebuilt image and allowing the host to read results directly from that memory. Its interpreter avoids using the Rust call stack for JavaScript calls, so execution can be suspended between instructions and resumed inside callbacks, getters, and asynchronous functions. A suspended tool call can be serialized as a continuation dump containing JavaScript code that recreates the run. A single compilation pass produces stack bytecode while erasing TypeScript during parsing, and clock, randomness, and weak references are supplied by the host so identical inputs produce byte-identical heaps in native and WebAssembly environments. djs is distributed as the @earendil-works/djs npm package and also provides a Rust CLI for running scripts, starting a REPL, and disassembling code. Its behavior is tested against dquickts, a QuickJS fork, and the repository includes documentation, architecture notes, testing instructions, and a pi codemode adapter.
dsh-superagent is a research-prototype proactive coding agent for a single GitHub repository, packaged for DeepSeek Harness. It listens to GitHub and optionally Feishu; when an authorized person requests work, it creates a work item, runs the task in its own agent session and directory, and replies in the original conversation, potentially with a pull request. A web panel shows items requiring human attention, while authorization rules limit when it acts and what it communicates. The repository provides documentation covering its philosophy, domain model, authorization, execution, prompts, panel, deployment, architecture, and adding sources. It is research code rather than an official DeepSeek product or production-ready release, and is licensed under MIT.
etcd is an open-source distributed key-value store for critical data in distributed systems. Written in Go, it uses the Raft consensus algorithm to maintain a highly available replicated log and exposes a user-facing gRPC API for key-value reads and writes. It supports automatic TLS with optional client-certificate authentication and includes the etcdctl command-line client, client libraries, server packages, and release binaries for macOS, Linux, Windows, and Docker. etcd is distributed under the Apache 2.0 license.
Herdr is a desktop application for terminals and coding agents, written in Rust on top of Ghostty's terminal engine. It organizes work into spaces, tabs, and panes, while an agents sidebar collects agents from those spaces and shows whether each is working, blocked, done, or idle. Its core keeps terminals running after the window closes; panes and spaces can run on local or SSH-connected machines, with remote work continuing while the local window is closed or the computer sleeps. The app supports layouts and JavaScript mods for customizing panels, menus, and dialogs, and provides command-line commands for inspecting the window and agent-driving guidance. Herdr is distributed for Linux, macOS, and Windows, with installers and source-build instructions in its repository. Release builds check for updates in the background and enable usage telemetry by default, which can be disabled in the privacy settings.
Kiri is a lightweight container runtime for macOS, Windows, and Linux, developed in Rust as an alternative to Docker Desktop and OrbStack. It runs Docker images and supports the Docker CLI, Compose, buildx, Docker login, amd64 images through Rosetta, and the Docker Engine API. On macOS and Windows it runs containers in a small Linux virtual machine; on Linux it can use the host kernel. Its guest agent starts containers directly with clone3, namespaces, and cgroups, without runc, a shim, or dockerd. Kiri also provides a native interface for containers, images, volumes, and networks, local .kiri.local domains with HTTPS, one-command single-node Kubernetes clusters, and Linux machines. It is distributed for macOS, Windows through WSL 2, and Linux, and is licensed under Apache-2.0; the guest kernel is GPL-2.0.
LivLiv Player is a lightweight browser library for rendering animated 2.5D characters with WebGL2. Its dependency-free core is about 30 KB gzipped; it loads LivLiv model packages, VRM avatars, and PNGTuber images through related APIs, with expressions, actions, and parameter controls. Optional browser modules provide lip sync from microphone, audio, or video and webcam face tracking through MediaPipe, with camera images processed on the device. The package is distributed through npm as @livliv-dev/player, and the source code is licensed under Apache License 2.0. Featured characters included in the repository have separate licensing and usage terms.
LZ4 is a lossless compression library and command-line utility implementing an extremely fast compression algorithm. It supports block and frame formats, streaming through multiple blocks, dictionary compression using the final 64 KB of an input dictionary, and dynamically tunable compression speed through an acceleration factor. LZ4_HC provides a higher-compression variant that uses more CPU time, while retaining the same decompression method and speed. The project includes C APIs for block, high-compression, and frame operations, along with build support for Make, CMake, Meson, and Visual Studio. The library is licensed under the BSD 2-Clause license; the command-line utility is licensed under GPL-2.0-or-later.
Mago Studio is an open-source, local-first desktop AI video editor for Windows, macOS, and Linux. Mago Director turns scripts into casts, shot lists, storyboards, and connected scenes; Mago Transform applies frame-accurate edits to selected video ranges, including masked edits, relighting, character replacement, and style transfer. Projects and video remain on the user's machine by default, with rendering through ComfyUI on a local GPU or through Mago Cloud when signed in. Cloud renders upload only the selected Shot, its source video, reference images, and masks. The application can install a bundled ComfyUI setup on Windows or connect to an existing ComfyUI installation; local rendering is optional, and cloud models are paid. The app is distributed under the GNU AGPLv3, while downloaded models retain their own licenses.
OmniDiff is a Rust command-line diffing tool by Marko Ivankovic that compares files according to their content. It uses tree-sitter grammars to structurally diff 25 programming and markup languages, showing changes in terms of functions, classes, structs, and other syntax nodes rather than only line-by-line text; unknown extensions fall back to plain-text comparison. It provides an interactive terminal viewer, plain-text headless output for scripts and CI, JSON output, and integrations with Git and Jujutsu. It also compares non-text content, including pictures pixel by pixel, archives member by member, fonts glyph by glyph, message catalogs message by message, and PDFs page by page. It can be installed from source with Cargo, through prebuilt releases and several package managers, or used through VS Code and Neovim integrations. The project is distributed under the GNU Affero General Public License version 3, with a separate commercial license available through GitHub Sponsors.
PR-Agent is an open-source, AI-powered code review agent and community-maintained legacy project of Qodo. It reviews and explains pull requests through commands such as `/describe`, `/review`, `/improve`, and `/ask`, and can find similar issues with `/similar_issue`. Its PR Compression strategy is designed to process both small and large pull requests, while JSON-based prompting supports customization of review categories and behavior. PR-Agent can run through a CLI, GitHub Actions, Docker, self-hosted deployments, or webhooks, and supports GitHub, GitLab, Bitbucket, Azure DevOps, and Gitea. It connects to models from OpenAI, Anthropic, Google, DeepSeek, Mistral, Ollama, and other providers reachable through LiteLLM. The repository is distinct from Qodo's hosted AI code review offering.
Rust is a systems programming language and toolchain developed by the Rust project. Its compiler, standard library, documentation, package manager and build tool Cargo, formatter rustfmt, linter Clippy, and editor support are maintained in the main repository. The language uses a rich type system and ownership model to enforce memory and thread safety at compile time, while targeting fast, memory-efficient software for critical services, embedded devices, and integration with other languages. Rust is primarily distributed under the MIT and Apache 2.0 licenses, with portions under BSD-like licenses.
sand3d is a Python library for writing 3D objects as short programs designed for LLMs and human readers. Built over Blender's bpy, it represents profiles, solids, edits, frames, joints, groups, and assemblies as named parts; parts, edges, and faces are addressed by name rather than hand-computed coordinates, and the same program produces the same mesh deterministically. It can export meshes and articulated assemblies to GLB, URDF, OpenUSD, OBJ, and STL for web viewers, game engines, robotics and simulation tools, CAD workflows, and 3D printing. The library is distributed under the Apache License 2.0 and includes example programs, documentation, and datasets of procedural reconstructions.
terminal-mac is a macOS virtualization tool for giving computer-use agents a separate macOS desktop on Apple silicon Macs. It runs a real macOS guest through Apple's Virtualization framework; a helper app in the guest captures the screen and receives input events, while a terminal viewer streams the display through Kitty, Ghostty, or iTerm2. Agents control the guest through an MCP server with tools for screenshots, clicking, dragging, scrolling, typing, keyboard input, and status checks, without moving the host Mac's pointer or changing its clipboard. The guest has separate apps, files, accounts, and permissions, with host-to-guest communication through shared control and tools volumes. The project supports multiple virtual Macs, lifecycle commands, configurable virtual CPUs, memory, disk, and display settings, and optional offline operation. It requires an Apple silicon Mac running macOS 14 or later; initial setup requires downloading a macOS restore image and completing macOS setup and privacy permissions inside the guest. The repository describes it as an early release candidate, with self-signed releases and no Apple notarization.
Zed is a high-performance, multiplayer code editor developed by Zed Industries, Inc., the creators of Atom and Tree-sitter. It is available for macOS, Linux, and Windows through direct downloads or local package managers; web support is not yet available. The source code is licensed primarily under GPL-3.0-or-later, with marked components under Apache-2.0.
Searchable transcript of Trending Open-Source GitHub Projects : Zed, claude-mem, Kiri, OmniDiff, PR-Agent & More #307 — ManuAGI - AutoGPT Tutorials (13:16). Search for a phrase, then click its timestamp to jump straight to that moment in the video.
Captions sourced from the original video on YouTube, published by ManuAGI - AutoGPT Tutorials. The video, its captions and all related intellectual property remain the property of their respective owners; AINotes claims no ownership. Provided for research, accessibility and search — see the Transcript Notice and Copyright Policy.
00:00 Every week, developers release powerful open-source GitHub projects. This is your weekly GitHub project update video. Top trending open- source GitHub projects. This week, you will discover useful and trending developer tools from fast editors and local AI video tools to coding agent helpers. Without wasting time, let's get started. >> Before we jump into today's project updates, here's a quick announcement for everyone.
00:24 We've launched a brand new YouTube channel called AI Agent Studio dedicated entirely to AI Agent projects, tutorials, and tools. So, if you're interested in staying up to date with the latest AI Agent open source projects, learning how to build your own agents, or exploring cuttingedge agent frameworks, make sure to check it out. Subscribe now to get weekly videos, in-depth guides, and realtime project breakdowns.
00:50 The link is right there in the description. Don't miss it. All right, let's get into today's video. >> Project number one, Zed. A fast multiplayer code editor in Rust. Zed is an open-source higherformance code editor built from scratch in Rust by the creators of Atom and Tree Sitter. It's engineered for speed, rendering on the GPU, so editing and scrolling stays smooth even in huge files.
01:13 Real-time collaboration is built in so you can code together and share a workspace. and it pairs AI assistance with your terminal language servers and Vim mode. It runs on Mac OS, Linux, and Windows. Install it and feel the speed. Project number two, NY Rust, the source of the Rust programming language. Naj is the official home of the Rust programming language, the open-source systems language built for performance, reliability, and memory safety without a garbage collector.
01:43 This repository holds it all. the Rust compiler, the standard library, and the cargo build tool and package manager. Its borrow checker guarantees memory and thread safety at compile time, and development happens in the open through the RFC process. It powers everything from operating systems to web services. Clone it and build the compiler. Project number three, Nangular.
02:06 Google's TypeScript framework for web apps. Angular is Google's open-source TypeScript first framework for building web applications from single page apps to large enterprise systems. It gives you a complete batteries included platform, a component model, a powerful router, forms and HTTP client, plus dependency injection and reactivity through signals, a CLI scaffolds, builds and tests your project and serverside rendering speeds first loads.
02:33 It's a full framework rather than a library with long-term support releases. Install the CLI and build your app. Project number four, Google Agent Skills. Google's official agent skills for Google Cloud. Agent Skills is Google's open-source repository of agent skills for its products with a heavy Google Cloud focus. Each skill is a portable skill MD following the open agent skills standard.
02:56 So, it grounds your agent in real Google expertise and through progressive disclosure loads only what's needed. skills cover Big Query, Cloud Run, Firebase, GKE, and the Gemini API, plus well architected security, reliability, and cost. It works with Gemini CLI, Cloud Code, Cursor, and more. Add them and build on Google Cloud. Project number five, LZ4, an extremely fast lossless compression library.
03:25 LZ4 is an open-source extremely fast lossless compression algorithm and library created by Yan Khaled. It trades some compression ratio for speed. It compresses and decompresses quickly with decompression especially fast, ideal for real-time use. A high compression mode, LZ4 headc squeezes files smaller while keeping fast decompression and a framing format packages streams for interoperability.
03:52 Written in portable C with a simple API, it's embedded everywhere. The Linux kernel, databases, file stems, and more. It's BSD licensed. Add it and compress it speed. Project number six, Nesh ESTD, a distributed key value store for your cluster. Nesh ESTD is an open-source distributed and reliable key value store for the most critical data in a distributed system.
04:15 A CNCF project best known as the backbone of Kubernetes. It uses the Raft consensus algorithm to replicate data across a cluster of machines so it stays consistent and available even when nodes fail. Written in Go, it offers watches, leases, and transactions over a simple API, making it ideal for configuration, service discovery, and leader election.
04:36 Run it and coordinate your cluster. Project number seven, Claude MEM, persistent memory across sessions for your agents. Claude MEM is an open-source plug-in that gives coding agents persistent memory across sessions. As you work, it quietly captures what the agent does. Uses AI to compress observations into summaries and stores them. Then injects the relevant ones into future sessions.
05:00 So Claude doesn't start from zero. Ask about past work and a mem search skill pulls it up. It keeps everything in a local SQLite database and works with Claude code, Codeex, Gemini, C-Pilot, and more. Install it and stop reexplaining. Project number eight, Deeptutor. An agentnative workspace for personalized learning. Deeptutor is an agentnative learning workspace that connects tutoring, problem solving, quizzes, research, and visualization in one system.
05:28 It offers a web app and a command line tool and runs locally through pip source or docker. Answers can draw on knowledge bases through several retrieval engines and it connects to model providers, MCP servers, and coding agents like claude code. Memory stays in readable files. It suits learners who want inspectable AI tutoring. Run it and start learning.
05:49 Project number nine, Perepper Agent. Open-source AI code review for pull requests. Perepper Agent is an open-source AI code review agent that works on pull requests. Commands like describe, review, improve, and ask run as comments on a PR or from the command line. Each using a single LLM call. It supports GitHub, GitLab, Bitbucket, Azure DevOps, and Gitia, and works with models from OpenAI, Anthropic, Google, and more through Light LLM.
06:16 It runs as a GitHub action in Docker or self-hosted. It suits teams automating PR reviews. Add it to a workflow. Project number 10, NEPS5. Port PS5 executables to Linux and Windows natively. Any PS5 is an open-source tool that automatically ports PS5 executables to run natively on Linux and Windows, not an emulator. A relinker rewrites the executable.
06:41 Any PS5 reimplements the PS5 system libraries it calls graphics, fonts, audio, save data, input, and a shader recompiler translates PS5 shaders to Vulcan. It's built for interoperability, research, and preservation. And it ships no copyrighted software, firmware, or keys. You bring your own executables. Written in C++, it's early and communitydriven.
07:06 Build it and port a title. Project number 11. Herder app. The desktop app for terminals and coding agents. Herder is the open-source desktop app for terminals and coding agents. The window onto the Herder runtime your agents run in. It arranges your terminals and agent sessions into layouts you can rearrange and customize so you watch many agents at once and switch what you see.
07:28 The app drives Herder Core which runs and coordinates the programs installing the core itself if it's missing. Built in Rust, it runs on Mac OS, Windows, and Linux. Install it and your agents. Project number 12, Terminal Mac. A Mac OS VM in your terminal for agents. Terminal Mac runs a Mac OS virtual machine inside your terminal, giving an agent its own Mac while you keep yours built on Apple's virtualization framework.
07:54 It downloads a restore image, boots a VM, and streams the guest desktop into your terminal. Computer use is builtin. An MCP server exposes screenshot, click, type, and key press with a CLI and skill. It plugs into cloud code. It's free, open-source, Apple silicon only, and early. Install it and give your agent a Mac. Project number 13. Player, a 2.5D web character player with lip sync.
08:21 Live Live Player is a lightweight open-source 2.5D character player for the web. It renders layered 2D art that turns its head with depth in WebGL2 with lip sync from a microphone or audio and webcam face tracking in the browser. One small API plays live, VRM, and PNG tuber characters alike with expressions and live control. Make a character from a single image on live.ai.
08:47 Then load its livive model. The player is Apache licensed. Install it and bring a character to life. Project number 14. Cope Anything. Point your agent at a game. Get co-op. Co-op Anything points your AI agent at any single player game to build co-op online or split screen for you. It's an open-source playbook plus an MCP server guiding an agent to detect a game's engine, reverse engineer it, write the networking, test it, syncs, and ship a mod.
09:16 A directory holds examples. Deis X, Dungeon Keeper, Fallout New Vegas, Ocarina of Time, and a Steam Relay means friends buy nothing extra. The agent asks before installing anything. Point it at a game and play together. Project number 15. Omniiff. Fast syntaxaware diffing for code and images. Omnidiff is a fast open-source syntax diff tool written in Rust.
09:43 Instead of comparing lines, it understands code structure. It classifies each change as an insert, delete, update, or move. pairs moved blocks with their counterpart and names the enclosing function or class matching renamed or reordered code across many languages. It even diffs images instead of saying binary file differs. Use it as a git diff tool or in VS Code and Neovim.
10:06 It's AGPL licensed. Install it and diff better. Project number 16. DSH Super Agent Duttoype, a proactive coding agent for one repository. DSH super agent is a proactive coding agent for one GitHub repository built as a bundle for Deep Seek Harness. It listens to GitHub and optionally FishU. When an allowed person asks for work, it opens a work item, works in its own session and replies in the same conversation, often with a pull request.
10:35 A web panel shows what needs a person. It is research code, not production software. It suits teams testing low attention agent workflows. Read the setup guide. Project number 17, Ciri Lightweight Docker Desktop replacement written in Rust. Kiri is a lightweight container runtime for Mac OS, Windows, and Linux. Written in Rust, it replaces Docker Desktop and Orbstack by running containers in one small Linux VM or directly on the kernel on Linux.
11:04 It speaks the Docker engine API, so the Docker CLI, Compose, and Buildex keep working. It also offers a one-command Kubernetes cluster, Linux machines, and a native window. It suits developers who want lighter local containers. Download it and run your first image. Project number 18, Meo Studio, local first AI video editing on your own GPU. Meo Studio is a local first desktop application for AI video editing available on Windows, Mac OS, and Linux.
11:34 It keeps projects and videos on the user's machine. MIGO Director turns a script into cast, shot lists, storyboards, and consistent scenes. Meo transform restyles footage frame by frame using masks and frame ranges. Renders run on a local GPU through Comfy UI or in Mo cloud after signin. It suits video creators who want local control. Download it and render your first shot.
12:00 Project number 19, DJs, a tiny JavaScript engine for AI scripts. DJs is a small JavaScript and TypeScript engine written in no std rust. It targets WessM32 and replaces QuickJS in Pi's code mode where model written scripts call tools and return data. All VM state lives in one bite arena. So a snapshot is a simple copy. Execution can suspend between any two instructions and resume later.
12:26 Results stay deterministic across native and Wessom builds. It suits developers running AI generated scripts. Try the CLI on a script. Project number 20. SAND 3D. Readable 3D modeling in Python for LLMs. SAN 3D is a Python library that lets LLMs write 3D objects as short programs built on Blender's BI. Programs describe profiles, solids, edits, joints, and assemblies.
12:50 And parts are reached by name instead of handcomputed coordinates. The same program builds the same mesh every time. One call exports GB, URDF, USD, OBJ or STL. The whole API fits in one prompt. It suits developers using LLMs for 3D and simulation assets. Install it with PIP. Thanks for watching. See you in the next update.