GitHub Trending Weekly Digest — Feb 11–15, 2026

I've been tracking GitHub Trending daily and decided to compile a weekly digest — deduplicated and ranked by how many days each project stayed on the chart. This week featured 13 unique projects across 5 days, with a strong AI + developer tooling theme throughout.

---

🏆 Dominated the Charts All Week

1. ChromeDevTools / chrome-devtools-mcp

⭐ Trending all 5 days (Feb 11–15) | 🔗 github.com/ChromeDevTools/chrome-devtools-mcp

What it does: Google's official MCP server for Chrome DevTools — lets AI coding assistants (Gemini, Claude, Cursor, Copilot) directly control and debug Chrome browsers.

Why it matters: AI coding agents can't see what's happening in the browser. This MCP server exposes Chrome DevTools' full capabilities to AI through a standard protocol — screenshots, network request monitoring, performance trace analysis, console logs with source maps, and Puppeteer-based automation.

Tech: TypeScript, Puppeteer, Chrome DevTools Protocol, MCP, CrUX API

---

📈 Trending Multiple Days

2. github / gh-aw

⭐ 2 days (Feb 11, 15) | 🔗 github.com/github/gh-aw

What it does: GitHub's official Agentic Workflows tool — write AI agent workflows in natural language Markdown and run them in GitHub Actions.

Why it matters: Traditional CI/CD requires complex YAML configs. gh-aw lets you describe tasks in plain language while AI agents execute them automatically, with multi-layered security (sandbox execution, input sanitization, network isolation, human approval gates).

Tech: Go, GitHub Actions, MCP Gateway, Agent Workflow Firewall, SHA-pinned dependencies

---

3. rowboatlabs / rowboat

⭐ 2 days (Feb 14, 15) | 🔗 github.com/rowboatlabs/rowboat

What it does: An open-source, local-first AI collaboration assistant that automatically builds knowledge graphs from emails and meeting notes.

Why it matters: Most AI tools "forget" everything between conversations. Rowboat maintains a persistent knowledge graph (Obsidian-compatible Markdown) that accumulates over time — like human memory that grows instead of resetting.

Tech: Electron, Obsidian Markdown vault, MCP protocol, Ollama/LM Studio, Deepgram voice, Gmail/Granola/Fireflies integration

---

4. tambo-ai / tambo

⭐ 2 days (Feb 12, 14) | Peak: 544 stars/day | 🔗 github.com/tambo-ai/tambo

What it does: A React generative UI SDK that lets AI agents directly render your React components.

Why it matters: Traditional AI chat can only output text — no interactive UI. Tambo lets you register components with Zod schemas, and the AI automatically selects and stream-renders the right UI (charts, kanban boards, forms). Say "show me sales data" and get a live <Chart>.

Tech: React, TypeScript, Zod, supports OpenAI/Anthropic/Gemini and more

---

5. SynkraAI / aios-core

⭐ 2 days (Feb 13, 14) | 🔗 github.com/SynkraAI/aios-core

What it does: An AI-driven full-stack development OS framework using multi-agent collaboration for automated software development.

Why it matters: AI coding assistants suffer from "lost context" and "inconsistent planning." AIOS solves this in two phases: ① Analyst/Architect agents collaborate on detailed PRDs and architecture docs; ② Scrum Master agent converts them into hyper-detailed dev stories so coding agents have full context.

Tech: JavaScript, Node.js (≥18), CLI-first architecture, SSE Dashboard, multi-agent system

---

6. google / langextract

⭐ 2 days (Feb 11, 12) | Peak: 1,654 stars/day | 🔗 github.com/google/langextract

What it does: Google's open-source Python library for extracting structured information from unstructured text using LLMs, with precise source attribution.

Why it matters: Extracting structured data from long documents (clinical notes, reports) while maintaining accuracy and traceability is hard. LangExtract supports chunked parallel processing, multi-pass extraction, and generates interactive HTML visualizations with highlighted source mapping.

Tech: Python, Gemini/OpenAI/Ollama, Pydantic, interactive HTML visualization

---

7. danielmiessler / Personal_AI_Infrastructure

⭐ 2 days (Feb 12, 13) | 🔗 github.com/danielmiessler/Personal_AI_Infrastructure

What it does: A personal AI infrastructure framework (PAI) designed to give everyone an AI system that amplifies their capabilities.

Why it matters: Most people lack the knowledge to build their own AI tool stack. PAI provides a complete install-configure-use framework with 23 feature packs and preset bundles, lowering the barrier to personal AI infrastructure.

Tech: TypeScript, Bun, Claude AI, modular Pack/Bundle architecture

---

8. microsoft / PowerToys

⭐ 2 days (Feb 11, 12) | 🔗 github.com/microsoft/PowerToys

What it does: Microsoft's official Windows productivity toolkit — 25+ utilities to customize and optimize the Windows experience.

Why it matters: Windows lacks many power-user productivity tools out of the box. PowerToys fills the gap with window management (FancyZones), quick launcher (PowerToys Run), batch rename, keyboard remapping, color picker, AI paste, and more.

Tech: C#, C++, WinUI 3, .NET, Windows API

---

📌 Single-Day Appearances

9. EveryInc / compound-engineering-plugin (Feb 11)

🔗 github.com/EveryInc/compound-engineering-plugin

Claude Code compound engineering plugin — Plan → Work → Review → Compound workflow. Invest 80% in planning and review, 20% in execution, achieving "compound interest" in engineering productivity.

Tech: TypeScript, Claude Code Plugin, Bun, multi-agent architecture

---

10. patchy631 / ai-engineering-hub (Feb 13)

🔗 github.com/patchy631/ai-engineering-hub

A collection of 93+ production-grade AI engineering projects covering LLMs, RAG, agents, and more — organized by difficulty level, each project runnable and deployable.

Tech: Jupyter Notebook, Python, LlamaIndex, Ollama, Streamlit

---

11. TelegramMessenger / MTProxy (Feb 13)

🔗 github.com/TelegramMessenger/MTProxy

Telegram's official MTProto proxy server with random padding for detection resistance, multi-key support, and Docker deployment — keeping Telegram accessible in restricted networks.

Tech: C, OpenSSL, MTProto protocol, Docker

---

12. nautechsystems / nautilus_trader (Feb 15)

🔗 github.com/nautechsystems/nautilus_trader

High-performance algorithmic trading platform with identical code for backtesting and live trading, eliminating strategy rewrite risk. Supports FX, equities, futures, options, and crypto.

Tech: Rust (core engine + tokio), Python (strategy layer), Cython, Redis, Docker

---

13. steipete / gogcli (Feb 15)

🔗 github.com/steipete/gogcli

A full-featured Google Workspace CLI — Gmail, Calendar, Drive, Contacts, Sheets, Docs and 15+ services from one command line. JSON output makes it perfect for AI agent integration.

Tech: Go, OAuth2, OS Keyring, JSON-first, Homebrew/AUR

---

💡 Takeaway

AI is eating developer tooling. From browser debugging (chrome-devtools-mcp) to CI/CD (gh-aw) to full-stack development (aios-core), the trend is clear — AI agents are becoming first-class citizens in the development workflow.

---

Data source: GitHub Trending | Compiled weekly by Tommy Zhang