📦 luwill/research-skills

Automate Academic Research with AI Agent Skills

Research workflows are being transformed by a collection of specialized AI agent skills designed to handle academic tasks with high precision. This project encapsulates common research processes into portable modules that integrate directly into Claude Code. It offers tools for writing systematic medical imaging literature reviews, generating professional academic slide decks from PDFs with automatic figure extraction, and drafting PhD-level research proposals. A collaborative multi-agent system also allows five distinct roles to work in tandem on survey papers. By adopting these structured workflows, you can replace manual drudgery with a more reliable and consistent research process.

🆔 @hackernewsgithubprojects

📦 jane-xiaoer/xiaoer-videolab

Xiaoer VideoLab: One-Click Local Video Downloads

Xiaoer VideoLab is the browser extension that finally lets you download videos directly to your local folder with a single click. Instead of relying on privacy-invasive extensions, this project uses a tiny, local background daemon powered by the battle-tested yt-dlp to handle your downloads. It supports over 1800 sites, including YouTube, Bilibili, and Instagram, without sending your data to external servers. Because all the processing happens right on your machine, you get full control over your files and keep your browsing private. Install the engine once and simplify your entire video-saving workflow for good.

🆔 @hackernewsgithubprojects

📦 saihgupr/dmgmaker

Create Professional macOS Installers with DMGMaker

Package your macOS applications into professional disk images using this dedicated creation tool. It simplifies the process by generating high-quality installers featuring live-rendered SwiftUI mesh gradients and a sleek glassmorphic interface that looks sharp on Retina displays. You can drag and drop your files directly into the intuitive interface or utilize the command-line support to integrate DMG creation into your automated build pipelines. By including specialized naming logic to prevent Finder display issues, this project ensures your installer looks polished and professional. Grab this tool to streamline your software distribution and create refined, user-ready installers effortlessly.

🆔 @hackernewsgithubprojects

📦 opensensenova/sensenova-skills

Automate Office Tasks with Modular AI Skills

The SenseNova Skills repository provides a modular toolkit for building AI-powered office assistants that automate complex productivity workflows. It operates by plugging into agent runtimes, allowing users to execute tasks like advanced Excel data analysis, deep research, and automated presentation generation. The project uses structured skill directories to define specific capabilities, such as cleaning Excel data, creating charts, or generating business-ready slide decks through standardized prompts. By integrating these tools, users can transition from raw data to finished reports or professional infographics without manual overhead. This project offers a practical way to standardize and scale AI-driven office automation.

🆔 @hackernewsgithubprojects

📦 tetherto/qvac

Build Peer-to-Peer AI Apps Locally with QVAC

QVAC enables you to run AI tasks like LLMs, speech recognition, and image generation entirely offline across mobile and desktop devices. By providing a unified JavaScript SDK, this open-source ecosystem removes the need for cloud providers, keeping data private and local. It also introduces unique peer-to-peer capabilities, allowing you to delegate complex AI inference to other machines. Whether you are building apps for Android, iOS, or desktop, this framework offers a consistent way to integrate high-performance models while supporting standard APIs. It is a powerful toolkit for developers who want to create truly decentralized, local-first artificial intelligence systems.

📰 https://news.ycombinator.com/item?id=48687796

🆔 @hackernewsgithubprojects

📦 blendi-remade/sprite-sheet-creator

Build Pixel Art Assets with AI

This application generates custom pixel art characters and game maps directly from text prompts using integrated artificial intelligence models. It streamlines the game development workflow by automatically creating multi-frame sprite sheets for side-scrolling or isometric RPG modes, complete with character animations and layered parallax backgrounds. Built with Next.js and the fal.ai platform, the tool also includes practical features like background removal, frame extraction, and interactive sandboxes for testing movement at adjustable speeds. It provides a specialized environment for creators to prototype 2D game assets efficiently, offering a cohesive pipeline from initial character generation to final animated preview.

🆔 @hackernewsgithubprojects

📦 zeroleaks/zeroleaks

ZeroLeaks: Autonomous AI Security Scanning

ZeroLeaks is the autonomous security scanner that systematically tests LLM systems for prompt injection and data extraction vulnerabilities. It utilizes a sophisticated multi-agent architecture featuring strategist, attacker, and evaluator agents to simulate real-world attacks using modern techniques like many-shot, crescendo, and chain-of-thought hijacking. The platform explores diverse attack vectors, including tool exploits and social engineering, while fingerprinting existing defenses to identify potential weaknesses before they can be exploited. By integrating research-backed methods into your workflow, this tool provides a robust way to assess your system's resilience. Secure your AI models by uncovering hidden vulnerabilities today.

🆔 @hackernewsgithubprojects

📦 mininglamp-ai/cider

Supercharge LLM Inference on Apple M5

Accelerate large language model inference on Apple Silicon by unlocking high-performance hardware capabilities that standard frameworks often leave behind. This project provides custom primitives that implement W8A8 and W4A8 quantization, effectively boosting prefill speeds by up to 1.9 times. It also introduces an optimized attention kernel inspired by FlashInfer, which delivers significant speedups during autoregressive decoding. By automatically patching your models to utilize these efficient Metal kernels, the toolkit ensures smoother and faster performance for your local AI workflows. Unlock faster token generation and more efficient memory usage on your Mac today.

🆔 @hackernewsgithubprojects

📦 moh4696/websites-100-audit

100 Free Tools That Should Be Illegal

This repository provides a highly curated directory of one hundred powerful and completely free digital tools spanning categories from media production to developer utilities. It solves the problem of finding reliable resources by vetting each link and offering transparent, honest summaries that clearly flag freemium models or legally grey content. Whether you need an in-browser Photoshop alternative, advanced academic search engines, or simple file compression, this collection acts as a vital reference guide for efficient productivity. By auditing these sites for availability, the author ensures that these quality tools remain accessible, helping you navigate the web with confidence.

🆔 @hackernewsgithubprojects

📦 formsmith746/sketchforge-3d

Build 3D Models in Your Browser with SketchForge

Design 3D models directly in your browser without needing complex software or cloud accounts. This local-first editor lets you create shapes, cut holes, and group parts to build geometry easily. You can import existing STL files into your workspace and export finished models in STL or OBJ formats. It even includes an AI scene bridge that allows you to drive the editor using Claude to push parametric shapes onto your workplane. It is a lightweight, efficient tool perfect for anyone wanting a simple way to sketch and prepare parts for fabrication. Start building your next design today.

🆔 @hackernewsgithubprojects

📦 zcai0612/up2you

Fast 3D Human Reconstruction from Photos

Generate detailed 3D human models from unconstrained collections of photos using this efficient, tuning-free reconstruction pipeline. Developed for ICLR 2026, the project simplifies the process of creating personal 3D avatars by automating complex stages like multi-view RGB generation, normal prediction, and shape parameter estimation. It utilizes advanced human modeling techniques to transform casual images into expressive 3D representations without requiring lengthy fine-tuning processes. Whether you are working with diverse datasets or personal photo archives, this repository provides the necessary tools and training configurations to achieve high-quality results. Explore this powerful approach to human digitization today.

🆔 @hackernewsgithubprojects

📦 opendcai/paper2any

Turn Research Papers Into Slides and Diagrams Automatically

Transform static research papers, text, or images into professional, editable presentation materials with a single click. This project provides a powerful workflow to generate technical architecture diagrams, academic posters, experimental plots, and full slide decks directly from your source documents. It simplifies the academic synthesis process by integrating intelligent layout tools, citation tracking, and rebuttal drafting support. Whether you need to turn a complex PDF into an editable presentation or convert a rough screenshot into a clean diagram, this tool handles the heavy lifting, helping you communicate your research findings with polished, professional visuals.

🆔 @hackernewsgithubprojects

📦 0xsero/ai-data-extraction

Extract Your Local AI Coding Assistant History

Personal AI coding assistants store massive amounts of interaction data locally on your computer. This repository provides a collection of Python scripts designed to discover and extract that history from tools like Cursor, Windsurf, Claude Code, and Trae. By scanning standard storage locations for SQLite databases and JSONL session files, the toolkit recovers complete conversation logs, including code context, multi-file selections, and suggested diffs. It serves as an automated way to gather your own data for personal machine learning projects or fine-tuning models. Always remember to scan your extracted files for sensitive secrets before using them elsewhere.

🆔 @hackernewsgithubprojects

📦 amazon-far/fpo-control

Flow Policy Gradients for Robot Control

This repository provides the official implementation of FPO++, a framework that advances robot control by replacing traditional Gaussian action distributions with learned flow models. By mapping noise to actions through an ordinary differential equation, this approach improves policy optimization for complex tasks. The project includes specialized setups for two core areas: velocity-conditioned locomotion for legged robots within the Isaac Lab simulation environment, and fine-tuning base policies across various manipulation tasks. Researchers can use these tools to train or refine robotic behaviors, leveraging provided scripts to manage environment interactions, evaluate performance, and experiment with different flow-based architectures for more effective robot learning.

🆔 @hackernewsgithubprojects

📦 ersinkoc/project-architect

Build Better Software with Project Architect

Project Architect is a documentation-first planning skill that turns rough project ideas into comprehensive, implementation-ready blueprints. By guiding you through an interactive discovery process, it generates five critical documents including specifications, implementation plans, task lists, and optional branding guides. This system identifies design patterns and recommends tech stacks with trade-off analyses, eventually synthesizing everything into a single-shot prompt for coding agents. Compatible with over forty agents like Claude Code and Cursor, it ensures your project starts with clear architecture rather than guesswork, allowing for efficient, structured development from the very first line of code.

🆔 @hackernewsgithubprojects

📦 pawandeep-prog/revpdf-release

Edit PDFs Privately and Offline with RevPDF

Edit your documents locally with RevPDF, a native tool designed for performance and complete privacy. By keeping all processing strictly on your device, this software ensures your sensitive files are never uploaded to a cloud server or tracked. You can easily split, merge, compress, and convert files between PDFs and high-quality images without needing an account or any internet connection. It offers a clean, fast experience for managing contracts and forms on Linux or macOS while avoiding bulky subscription-based alternatives. Simplify your document workflow today by switching to a reliable, offline-first editor that puts your data security first.

🆔 @hackernewsgithubprojects

📦 pluviobyte/video-production-skills

Master Reusable AI Video Production with This Professional Library

This repository provides a professional collection of reusable AI video production skills designed for creators and developers who rely on AI agents to streamline their workflows. Rather than serving a single project, it acts as a modular library that covers essential production stages including video recreation, motion design, and high-quality quality assurance. You can leverage specific tools for tasks like creating cinematic dark-themed product shorts, generating clean black-and-white text openers, or performing evidence-driven reference video analysis. These skills are fully compatible with common AI coding environments and agents, offering a structured, reliable way to automate your video production process.

🆔 @hackernewsgithubprojects

📦 researai/autofigure

AutoFigure: AI-Powered Scientific Illustration Generation

AutoFigure is the research tool that turns complex scientific descriptions and academic papers into publication-ready visual diagrams. By leveraging a dual-agent system that performs iterative generation and evaluation, it transforms raw text into high-quality SVG or XML outputs compatible with common diagramming software. Users can extract methodology directly from PDFs or provide natural language prompts to create custom visuals. With integrated AI-powered enhancement features, it offers an efficient way to beautify your work while maintaining scientific rigor. Whether you are drafting a paper or building a presentation, this system simplifies the visual communication of technical concepts.

🆔 @hackernewsgithubprojects

📦 unconv-ai/un-0

Generating Images with Coupled Oscillators

Create images by integrating the phase dynamics of a population of coupled oscillators rather than relying on traditional diffusion or adversarial methods. This project explores how dynamical systems function as a specialized computing substrate, potentially offering a significantly more energy-efficient alternative to standard digital AI accelerators. The repository provides a reference implementation in PyTorch, complete with training pipelines for CIFAR-10 and ImageNet-64, alongside pretrained checkpoints that you can use to generate samples immediately. This approach demonstrates a unique, physics-inspired path for future machine learning hardware, showing that alternative computational models can produce high-quality results while rethinking standard AI architecture.

🆔 @hackernewsgithubprojects

📦 abundantbeing/hermes-browser-extension

Connect Your Browser to Hermes Agent

Hermes Browser Extension is the side panel tool that bridges your active web context to the Hermes Agent runtime. This browser-native extension communicates directly with your local or remote Hermes API server, allowing you to seamlessly integrate your models, tools, and memory into your daily workflow. It captures readable page text, metadata, and links, wrapping them as untrusted context before sending them to your configured sessions. With features like voice dictation and a sleek, themeable interface, this alpha tool lets you harness the power of your existing Hermes setup to analyze and interact with content right inside your browser.

🆔 @hackernewsgithubprojects