GitHub Trending Projects: Daily (2026-11-09)

This issue includes 9 trending open-source projects, adding a total of ⭐ 7,172 stars. Popular languages: Python, TypeScript.

Data source: GitHub Trending | Statistics period: Daily | Updated: 2026-11-09

📋 Overview

Rank	Project	Language	⭐ Total Stars	New This Period
1	sherlock	Python	103,537	+2,373
2	nvke	Go	854	+345
3	OpenMausBot	TypeScript	18,264	+1,439
4	claude-quickstarts	Python	19,424	+32
5	agent-browser	Rust	47,609	+815
6	astryx	TypeScript	16,806	+875
7	skills	Python	189,658	+792
8	mujoco	C++	16,420	+41
9	billion-context	TypeScript	5,157	+460

---

🔍 Project Details

1. sherlock-project/sherlock

Find social media accounts by username across social networks

Sherlock is a command-line tool written in Python that finds corresponding accounts on numerous social networks based on a username. It supports searching one or multiple usernames at a time and can save results as text, CSV, or XLSX files; it also offers parameters for proxies, timeouts, searching only specified sites, opening results in a browser, and more. Suitable for people who need to confirm which platforms a username is registered on or conduct account investigations.

Language: Python · Stars: 103,537 · Forks: 12,131 · New this period: +2,373

2. dsx-ai-factory/nvke

NVIDIA Kubernetes Engine (NVKE), a management plane for bare-metal GPU Kubernetes.

NVKE is a management plane for platform teams, installed on your own NVIDIA GPU infrastructure, used to create and operate multiple Kubernetes workload clusters on demand. Each workload cluster has its own control plane, etcd, and node pools, accessible via the Kubernetes API and kubectl. It is currently marked as Preview and not recommended for production; the infrastructure provider currently supports only NICo. It also handles node identity attestation, etcd encryption, and cluster lifecycle management, but upgrades, backup recovery, and other management-cluster functions remain limited.

Language: Go · Stars: 854 · Forks: 111 · New this period: +345

3. milind-soni/OpenMausBot

Open-source Grok Bot alternative with VMs for bots.

OpenMausBot is an open-source chat-based agent application inspired by Grok Bot. Each sidebar bot is a locally running Claude, Codex, or Grok agent, with independent models, cloud or local VMs, and apps like Gmail, Slack, GitHub connected via Composio. It is local-first: agent processes, records, and keys are stored on the machine, and it supports approval cards, channels, triggers, voice, and MCP servers. Desktop versions are available for macOS, Windows, and Ubuntu. Suitable for users who want to bring their own model accounts and manage multiple AI agents locally.

Language: TypeScript · Stars: 18,264 · Forks: 2,534 · New this period: +1,439

4. anthropics/claude-quickstarts

A collection of projects to help developers quickly start building deployable apps with the Claude API.

This repository brings together multiple Claude API quick-start projects, each usable as a base for customization. Examples include customer service agents, financial data analysis, computer use demos, browser automation, autonomous coding agents, and various Managed Agents-based chat, scheduled briefings, knowledge base, Slack, Linear, Sentry, and more. Requires an Anthropic account to obtain an API key. Suitable for developers who want to quickly build deployable apps on Claude.

Language: Python · Stars: 19,424 · Forks: 3,280 · New this period: +32

5. vercel-labs/agent-browser

Browser automation CLI for AI agents.

agent-browser is a native CLI written in Rust that provides a command-driven browser for AI agents. It reads pages as compact text with element reference markers via snapshots, allowing agents to click, fill, and read on those references, while keeping the browser open between commands. It controls the browser via a background daemon and the Chrome DevTools Protocol, and supports sessions, selectors, waits, screenshots, MCP servers, and skills. Suitable for scenarios where an AI operates web pages from the shell.

Language: Rust · Stars: 47,609 · Forks: 3,258 · New this period: +815

6. meta/astryx

A fully customizable, agent-ready open-source design system.

Astryx is an open-source design system from Meta, built on React 19+ and StyleX, currently in Beta. It offers 150+ accessible components, brand-level theming, dark mode, templates, and a CLI. Components are open internally, and you can use swizzle to eject full source into your project, or override styles with Tailwind, CSS Modules, or plain CSS. Theming is customized via CSS custom properties without forking or wrapping components. Suitable for teams building React apps where humans and AI assistants share the same set of components and documentation.

Language: TypeScript · Stars: 16,806 · Forks: 1,322 · New this period: +875

7. anthropics/skills

Agent Skills public repository.

This is Anthropic's repository for Agent Skills examples and specifications. A Skill is a folder containing SKILL.md instructions, scripts, and resources that Claude can load dynamically to complete specific tasks. The repo includes examples in creative design, development techniques, business communication, and document processing, and open-sources the source code for docx, pdf, pptx, xlsx document skills; it also provides the Agent Skills specification and templates. It can be registered as a Claude Code plugin marketplace, and some skills are already available in Claude.ai and the API. Suitable for developers learning or creating custom skills.

Language: Python · Stars: 189,658 · Forks: 22,349 · New this period: +792

8. google-deepmind/mujoco

Multi-Joint dynamics with Contact (MuJoCo), a general-purpose physics simulator.

MuJoCo is a general-purpose physics engine maintained by Google DeepMind, aimed at robotics, biomechanics, graphics animation, machine learning, and more, providing fast and accurate simulation of articulated structures and environment interactions. It offers a C API and preallocated low-level data structures, built-in desktop and browser visualization, Python bindings, and a Unity plugin; it can be installed via precompiled binaries or PyPI. Suitable for researchers and developers needing physics simulation, machine learning, or motion control research.

Language: C++ · Stars: 16,420 · Forks: 1,953 · New this period: +41

9. ranxianglei/billion-context

Context compression plugin supporting month-long single sessions (billions of tokens), small context window (100K is enough), and token savings (usage drops to about one-fifth). billion-context is all you need

billion-context is a context compression plugin/proxy that runs between an agent and the model API. It rewrites Anthropic/OpenAI streaming responses and injects tools like compress, decompress, search_context, and acp_status, letting the model decide when to compress consumed conversations into hierarchical summaries. The README claims it can reduce context window requirements to about 100K and token consumption to about one-fifth, and supports multiple coding clients such as pi, opencode, Claude, and Codex, with integration via native plugins, launchers, and URL prefixes. Suitable for long coding sessions and developers who want to save tokens.

Language: TypeScript · Stars: 5,157 · Forks: 406 · New this period: +460

---

This article is automatically aggregated by AI225 GitHub Trending, covering 9 trending open-source projects. Content is based on public GitHub Trending data; Chinese translation and interpretation are AI-generated, may contain omissions or errors, and are for reference only. For the full list and original descriptions of each project, visit: https://gh.ai225.com/daily/2026-11-09