1. reverse-skill by zhaoxuya520 - Reverse Engineering / Authorized Penetration Testing / Security Research Skill Router Pack
Language: PowerShell · Stars: 20289 · +10091 this week · View on GitHub →
What it is. A PowerShell-based skill router pack for AI coding clients (Claude Code, Kiro, Cursor, Cline) that routes reverse engineering, authorized penetration testing, and security research tasks to scenario-specific methodologies, toolchains, and scripts.
Why it's trending. According to the repository metadata it gained over 10,000 stars in the past week, and the README highlights a consolidated routing framework that turns scattered reverse-engineering and offensive-security playbooks — covering APK, mobile, binary, JS, malware, CTF, firmware, pwn, EDR bypass, and more — into repeatable workflows an AI agent can dispatch instead of guessing commands.
Who should care. Reverse engineers, security researchers, CTF competitors, and authorized pentesters who use AI coding agents and want structured, scenario-aware guidance for tasks like binary analysis, JS deobfuscation, mobile reversing, red-team orchestration, and similar offensive-security workflows.
2. AI-For-Beginners by microsoft - 12 Weeks, 24 Lessons, AI for All!
Language: Jupyter Notebook · Stars: 62989 · +9164 this week · View on GitHub →
What it is. AI-For-Beginners is a 12-week, 24-lesson beginner-friendly curriculum from Microsoft that teaches Artificial Intelligence concepts through practical lessons, quizzes, and labs covering TensorFlow, PyTorch, and AI ethics.
Why it's trending. The repository gained over 9,000 stars this week, attracting interest with a curriculum spanning symbolic AI, neural networks, deep learning, and genetic algorithms that is now accessible in over 50 languages via automated translations.
Who should care. Beginners and self-learners seeking a structured introduction to AI—including knowledge representation, deep learning with TensorFlow and PyTorch, and neural architectures for images and text—would benefit most from this multi-language curriculum.
3. TencentDB-Agent-Memory by TencentCloud - TencentDB Agent Memory is a team-level memory hub for AI Agents
Language: TypeScript · Stars: 17194 · +6444 this week · View on GitHub →
What it is. A team-level memory hub for AI agents, written in TypeScript, that converts conversations, documents, and code into four reusable memory assets (Chat Memory, Skill, LLM-Wiki, and Code-Graph) that can be shared across agents and frameworks.
Why it's trending. The repository added 6,444 stars this week and recently shipped a v2.0.0+ release alongside a data-migration tool from earlier versions and an actively evolving Team Memory Beta.
Who should care. Developers and small teams running multi-agent workflows on tools like Claude Code or CodeBuddy who want persistent, portable context, skills, wikis, and code-indexed knowledge shared across agents rather than rebuilt each session.
4. kaneo by usekaneo - 🎯 All you need. Nothing you don't.
Language: TypeScript · Stars: 7556 · +3078 this week · View on GitHub →
What it is. Kaneo is an open source, self-hosted project management platform built in TypeScript and licensed under MIT.
Why it's trending. It attracted 3,078 stars this week, drawing attention for its minimalist philosophy, self-hosted design, and built-in MCP (Model Context Protocol) server that lets AI tools manage tasks and projects.
Who should care. Software teams looking for a lightweight, self-hostable project management tool with Docker, Kubernetes, and AI assistant integrations.
5. airllm by lyogavin - AirLLM 70B inference with single 4GB GPU
Language: Jupyter Notebook · Stars: 29786 · +5222 this week · View on GitHub →
What it is. AirLLM is an open-source inference library that enables running extremely large open-source large language models on low-memory consumer hardware, including 4GB GPUs, without requiring quantization, distillation, or pruning.
Why it's trending. It is trending this week following its July 2026 update that added support for Kimi K3, the largest open-source model released to date, which runs on under 4GB of VRAM via per-expert streaming for sparse Mixture of Experts models, and the repository has gained over 5,200 new GitHub stars in the past week.
Who should care. It is intended for AI developers, researchers, and hobbyists who want to run large, state-of-the-art open-source language models on low-end consumer hardware (such as entry-level GPUs or Mac devices) without needing to perform model compression or access high-end enterprise compute.