1. luongnv89/claude-howto Not new today
A visual, example-driven guide to Claude Code — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
First seen: March 30, 2026 | Streak: 2d
Analysis
The `luongnv89/claude-howto` repository provides a comprehensive, project-based educational guide designed to help developers master Claude Code beyond basic prompt interactions. It moves past simple reference documentation by offering a structured, 10-module learning path that covers advanced features such as slash commands, custom hooks, MCP servers, and multi-agent orchestration. Technically, the guide utilizes visual Mermaid diagrams, interactive CLI-based self-assessments, and a library of production-ready, copy-paste templates that allow users to integrate automated workflows directly into their local development environments.
This project is an invaluable resource for developers and engineering teams who want to significantly accelerate their productivity by transforming Claude Code from a simple chatbot into a sophisticated, agent-driven development assistant. It is trending because it bridges the gap between official technical documentation and real-world implementation, saving users hours of manual configuration through its highly practical, example-driven format. By providing a clear roadmap for everything from basic slash commands to complex automated pipelines, it empowers users to unlock the full potential of their AI tools with minimal friction.
2. microsoft/VibeVoice Not new today
Open-Source Frontier Voice AI
First seen: March 29, 2026 | Streak: 3d
Analysis
VibeVoice is an open-source research framework from Microsoft that provides a suite of advanced voice AI models for both Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) applications. The project utilizes an innovative architecture featuring continuous acoustic and semantic speech tokenizers operating at an ultra-low 7.5 Hz frame rate, combined with a next-token diffusion framework that integrates Large Language Models (LLMs) to ensure contextually coherent output. Key features include the ability to process or generate long-form audio spanning up to 90 minutes in a single pass, native support for multilingual speech, and real-time streaming capabilities via the 0.5B parameter model.
Researchers, audio engineers, and AI developers focused on long-form conversational content, such as podcasts or multi-speaker dialogue, would benefit most from these high-fidelity, unified voice models. The project is trending because it bridges the gap between massive-scale speech processing and computational efficiency, offering a robust alternative for creating naturalistic, long-duration synthetic speech and structured, timestamped transcriptions. Furthermore, by providing fine-tuning capabilities and vLLM inference support, it lowers the barrier for the open-source community to build sophisticated, domain-specific voice applications while prioritizing modularity and performance.
3. Yeachan-Heo/oh-my-claudecode Not new today
Teams-first Multi-agent orchestration for Claude Code
First seen: March 26, 2026 | Streak: 3d
Analysis
*Oh-my-claudecode* (OMC) is a sophisticated multi-agent orchestration framework designed to enhance Claude Code by automating complex development workflows through a "team-first" approach. Technically, it functions as a plugin that manages staged pipelines—handling planning, execution, verification, and automated fixing—while leveraging specialized agents and smart model routing to optimize for both performance and token costs. It further extends capabilities by integrating tmux-based CLI workers, allowing users to spawn and orchestrate parallel processes across different models like Codex and Gemini directly within their terminal.
This project is ideal for software engineers and development teams looking to scale their productivity by offloading routine coding, debugging, and architectural tasks to a highly autonomous, multi-model system. It is trending because it drastically lowers the barrier to entry for complex AI-assisted development, offering a "zero-learning-curve" interface that replaces manual prompts with intelligent, persistent, and self-correcting workflows. By turning vague requirements into structured, verifiable code through deep interviewing and automated feedback loops, OMC effectively transforms the command line into a collaborative, multi-agent development studio.
4. shanraisshan/claude-code-best-practice Not new today
practice made claude perfect
First seen: March 16, 2026 | Streak: 2d
Analysis
The `claude-code-best-practice` repository serves as a comprehensive knowledge base and implementation guide for optimizing the usage of Claude Code. It provides a structured framework for managing agentic workflows through custom commands, subagents, persistent memory files (CLAUDE.md), and Model Context Protocol (MCP) servers. Technically, it leverages hierarchical configurations and modular skill injections to orchestrate complex development tasks, allowing users to move beyond basic prompts into autonomous, multi-agent coordination and sophisticated task automation.
This project is highly beneficial for power users, software engineers, and AI architects who want to transition from casual LLM interactions to a professionalized, agentic development environment. It is currently trending because it aggregates decentralized expert wisdom—including workflows from creators like Boris Cherny—into a single actionable source for mastering modern "vibe coding" and AI-assisted engineering. By offering standardized patterns for context management and agent delegation, the repository addresses the common challenges of AI tool bloat, helping developers maintain focus and code quality in increasingly automated workflows.
5. NousResearch/hermes-agent Not new today
The agent that grows with you
First seen: March 01, 2026 | Streak: 1d
Analysis
The Hermes Agent is a self-improving AI framework developed by Nous Research that functions as a persistent, autonomous assistant capable of evolving alongside its user. Technically, it utilizes a unique "learning loop" that allows the agent to synthesize new skills from experience, store long-term memories across sessions, and integrate with over 200 different LLM providers through a flexible, modular architecture. The project supports seamless deployment across varied environments, including low-cost VPS setups, cloud clusters, and serverless infrastructure, while offering a robust toolset that includes messaging gateways, task scheduling, and MCP integration.
This project is primarily designed for power users, developers, and AI enthusiasts who require a highly customizable, persistent assistant that grows more capable through continuous usage. It is currently trending because it addresses the growing demand for autonomous agents that transcend simple chat interfaces to provide genuine, context-aware utility and skill acquisition. By enabling users to maintain a cohesive digital identity and persistent workflow across diverse platforms and LLM models, Hermes Agent offers a compelling, open-source alternative to locked-in proprietary AI ecosystems.
6. obra/superpowers Not new today
An agentic skills framework & software development methodology that works.
First seen: February 04, 2026 | Streak: 1d
Analysis
Superpowers is an agentic framework designed to standardize the software development lifecycle by integrating a library of composable, automated "skills" directly into coding agents like Claude Code and Cursor. Instead of jumping straight to code, the system enforces a structured workflow that begins with Socratic brainstorming, proceeds to detailed implementation planning, and culminates in a rigorous subagent-driven development cycle. Technically, it functions as a plugin that triggers mandatory engineering protocols—such as strict test-driven development (TDD), git worktree management, and iterative code reviews—ensuring that agents maintain high-quality standards and consistent progress without requiring manual intervention.
This project is highly beneficial for developers and software teams looking to minimize the "hallucination" and ad-hoc coding patterns often associated with AI-assisted programming. It is trending because it addresses a critical gap in the current AI developer tool ecosystem: the transition from simple code generation to reliable, autonomous project execution that mirrors professional engineering discipline. By embedding best practices like YAGNI and DRY into the agent's core decision-making loop, Superpowers provides a repeatable, scalable methodology for building robust software applications.
7. microsoft/agent-lightning Not new today
The absolute trainer to light up AI agents.
First seen: February 02, 2026 | Streak: 1d
Analysis
Agent Lightning is a versatile Microsoft Research framework designed to optimize AI agents through advanced techniques like reinforcement learning, automatic prompt optimization, and supervised fine-tuning. It stands out by allowing developers to integrate these training capabilities into existing workflows with virtually no code changes, regardless of the underlying agent framework used (such as LangChain, AutoGen, or CrewAI). Technically, the system utilizes a lightweight tracer to capture structured spans of agent behavior—including prompts, tool calls, and rewards—which are stored in a central hub and processed by optimization algorithms to iteratively refine prompt templates or policy weights.
Researchers and AI engineers building complex, multi-agent systems will benefit most from this tool, as it provides a scalable, framework-agnostic solution for continuous model improvement. The project is currently trending because it addresses the significant pain point of "training friction," enabling developers to scale agent performance without undergoing massive architectural rewrites or becoming locked into specific ecosystems. By offering a proven path to stable, large-scale training—as demonstrated in community projects achieving up to 128-GPU scaling—it empowers teams to move from basic prototyping to production-grade, self-correcting agents with ease.
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
First seen: March 31, 2026 | Streak: 1d
Analysis
PaddleOCR is a versatile, industry-leading toolkit designed to transform complex documents, PDFs, and images into structured, LLM-ready data formats like Markdown and JSON. Its technical prowess relies on the high-performance PP-OCRv5 recognition models and the recently released PaddleOCR-VL-1.5, a 0.9B parameter vision-language model that excels at interpreting warped, scanned, or skewed documents. By combining a dynamic high-resolution visual encoder with lightweight language models, the framework provides state-of-the-art accuracy across 100+ languages while maintaining the resource efficiency required for edge and cloud deployment.
This project is a critical asset for developers and enterprises building intelligent RAG (Retrieval-Augmented Generation) pipelines, AI agents, and automated data extraction systems. It has gained significant traction because it bridges the gap between messy, real-world visual documents and the structured input requirements of Large Language Models, making it the preferred choice for ecosystem tools like Dify and RAGFlow. With its robust support for diverse hardware backends and its ability to handle complex layout challenges, PaddleOCR serves as an essential, production-ready foundation for any organization looking to scale their document-to-AI intelligence workflows.
My Codex Skills
First seen: March 31, 2026 | Streak: 1d
Analysis
Dimillian/Skills is a collection of modular, reusable engineering agents designed to be integrated into the Codex development environment to automate recurring tasks on Apple platforms. By placing these self-contained skill folders into a designated local directory, developers can leverage specialized agents for complex workflows, including iOS debugging, App Store changelog generation, and Swift concurrency analysis. The project emphasizes a multi-agent approach, utilizing automated swarms to handle code reviews, bug hunts, and large-scale refactoring efforts with systematic, prioritized feedback.
This repository is particularly valuable for professional iOS and macOS developers who want to scale their productivity by offloading repetitive maintenance and architectural tasks to specialized AI agents. It is gaining traction because it addresses the modern challenge of managing increasingly complex SwiftUI and Swift 6 concurrency requirements through structured, actionable automation. By providing a standardized framework for sharing these "skills," the repository empowers engineering teams to implement consistent coding standards and rapid diagnostic workflows across their entire project ecosystem.
10. sherlock-project/sherlock Not new today
Hunt down social media accounts by username across social networks
First seen: March 30, 2026 | Streak: 2d
Analysis
Sherlock is an open-source command-line tool designed to identify the presence of a specific username across hundreds of different social media platforms. By utilizing a collection of unique URL structures for each site, the application sends HTTP requests to verify if a profile exists and parses the server responses to confirm matches. This automated process replaces the tedious manual effort of searching for an individual's digital footprint, allowing users to consolidate OSINT (Open Source Intelligence) data efficiently.
This project is highly valuable for cybersecurity professionals, investigative journalists, and researchers who need to verify identities or map an individual’s online presence during background checks. It is trending because of the increasing importance of digital identity verification and the tool's accessibility, which allows users with varying levels of technical expertise to perform complex reconnaissance tasks. As data privacy concerns rise, Sherlock provides a transparent and powerful utility for those looking to understand the breadth of information publicly available across the internet.