📦 hilbertraumai/hilbertraum

HilbertRaum: Your Offline AI Workspace

HilbertRaum is the private AI workspace that keeps every byte of your data strictly on your computer. It solves the anxiety of sending sensitive documents to the cloud by running powerful language models entirely offline, using a curated list of open-weight models that it automatically matches to your hardware. The standout feature is its portable, encrypted design, which allows you to place the entire app, including all models and your private library, on a single USB drive. You can simply plug it into any machine and work with your documents, knowing that nothing is ever transmitted, tracked, or stored remotely.

🆔 @hackernewsgithubprojects

📦 merlinhu1/codex-game-studio

Codex Game Studio

Imagine turning a standard AI chat into a fully structured game studio that lives right on your computer. Codex Game Studio does exactly that by scaffolding fifty distinct agents, ninety skills, and sixty-one workflows directly into your project files. The coolest part is that everything is visible in plain text, so you can review and version control your entire virtual team's decisions without relying on a cloud dashboard. It keeps your AI assistant honest by enforcing strict validation rules, ensuring that every creative choice is documented and traceable. This is the perfect way to build serious games without the chaos of lost chat history.

🆔 @hackernewsgithubprojects

📦 crackfireflyorder/football-manager-2025-nexus

Football Manager 2025 Nexus: A Windows Workspace for Football Managers

Football Manager 2025 Nexus is the Windows workspace that finally keeps your planning, documents, and team updates in one clean spot. Instead of juggling separate tools for scheduling and editing, it bundles a rich text editor, a visual analytics dashboard, and a calendar with recurring events into a single application. This setup removes the clutter of switching between apps, letting you focus on your Football Manager 2025 tasks without the usual digital noise. It is designed to be fast and simple, ensuring your team workload and deadlines are always visible at a glance. If you want a tidy, all-in-one hub for your management workflow, this is the tool to try.

🆔 @hackernewsgithubprojects

📦 pentagonconure42/youtube-downloader-zen

The Truth About Youtube Downloader Zen

Youtube Downloader Zen is the video editing toolkit that hides a major red flag behind a simple interface. While it advertises professional timeline editing, GPU acceleration, and cinema-ready export presets, the repository itself contains no actual code. It is purely a text file describing a software package you are told to download from an external website. This disconnect is the real story here. Instead of open source transparency, it directs users to a binary release from a third-party domain. The project serves as a cautionary tale about verifying sources before installing heavy tools. Always inspect the code before you trust the promise.

🆔 @hackernewsgithubprojects

📦 kgoedecke/doop

Doop: Design With AI Live

Doop lets you design alongside AI agents in real time, right on a shared canvas. It is an open source alternative to Paper.design where you and your bots build together, live. Connect a tool like Claude Code, and watch it sketch out frames, stream the HTML in chunks, and even review its own work with screenshots while you watch. You see its cursor, its status, and every edit as it happens, making it feel like a true team. If you want to stop prompting and start collaborating with your code, Doop is the place to try.

📰 https://news.ycombinator.com/item?id=49411550

🆔 @hackernewsgithubprojects

📦 aleksandarskrbic/khaos

Khaos: Break Your Kafka on Purpose

You can intentionally kill your message brokers while traffic is still pouring in, and Khaos does exactly that to save your production system from itself. This open source tool generates realistic fake data and simulates heavy producer and consumer traffic so you can find weak spots before your real users do. The coolest part is that you can run specific chaos scenarios, like stopping brokers or forcing consumer lag, directly from simple text files without writing a single line of code. It handles the messy stuff like authentication and different data formats, so you can focus on seeing where your system actually breaks.

🆔 @hackernewsgithubprojects

📦 zai-org/glm-asr

GLM-ASR: The Whisper Killer for Quiet Rooms

GLM-ASR is the open-source speech recognition model that finally listens to the quiet stuff. Most AI voice tools struggle when you mumble, whisper, or speak in a noisy meeting, but this one was built specifically to catch those low-volume sounds that usually get lost. It packs impressive accuracy into a compact package, handling everything from standard Mandarin to Cantonese dialects with ease. If you have been dealing with transcriptions that keep missing your softer moments, this is the tool that actually hears you, turning tricky audio into clear text without the usual guesswork.

🆔 @hackernewsgithubprojects

📦 affaan-m/agentshield

AgentShield: The Security Scanner for AI Agents

AgentShield is the security scanner that protects your AI agent configurations from hidden threats. It digs into your setup to catch critical issues like hardcoded secrets, overly permissive tool access, and risky server connections that could lead to data leaks or system breaches. Instead of leaving your AI setup vulnerable, this tool provides a graded report with actionable fixes, ensuring your automation is safe before it runs. It is an essential safety net for anyone building with AI agents.

🆔 @hackernewsgithubprojects

📦 open-gsd/gsd-pi

GSD PI Keeps Coding Agents on Task

An AI coding agent can lose track of the big picture after just a few minutes of autonomous work. GSD PI solves this by giving agents a durable memory system that keeps them aligned with the original project goals for hours. It acts as a persistent guide, ensuring long-running tasks stay focused and accurate without human intervention. This is useful for developers who want reliable, uninterrupted code generation. The takeaway is simple: GSD PI gives AI coding agents the long-term memory they need to work independently and effectively. It is a practical tool for anyone looking to automate complex coding workflows with precision and stability.

🆔 @hackernewsgithubprojects

📦 unitshopkeeper56/cinema-4d-flow

Cinema 4D Flow: The Silent Powerhouse

Cinema 4D Flow is the Windows toolkit that handles heavy architectural 3D work without slowing down your machine. It solves the common problem where professional rendering software feels bloated and sluggish, offering a streamlined experience that stays out of your way. The standout feature is its high-performance engine, which is optimized to chew through complex scenes and massive assemblies while keeping resource usage light. This means you can focus on your design details instead of watching a progress bar. It supports all the major industry file formats, so it slots right into your existing workflow without forcing you to change habits.

🆔 @hackernewsgithubprojects

📦 cosmicstack-labs/mercury-agent-skills

Mercury Agent Skills: The Missing Manual for Your AI

Mercury Agent Skills is a collection of over a hundred hand-written playbooks that teach AI agents how to do real developer work. Each skill is a simple, plain-text file that slots directly into your coding assistant, giving it expert-level context on tasks like debugging, security checks, or frontend design without needing custom code. This solves the problem of generic AI responses by providing specific, tested instructions that keep your agent focused and efficient. It is essentially a shared library of professional best practices that any team can use to make their AI tools more reliable and useful for daily coding tasks.

🆔 @hackernewsgithubprojects

📦 disler/inkwell-agent-sandboxes-and-software-factory

Inkwell: The Self-Running Software Factory

Inkwell is a minimalist blog app that acts as a test payload for a much wilder idea: a software factory that builds code without you touching the keyboard. The system spins up a disposable virtual machine in about ten seconds, installs the app and a set of AI agents inside it, and lets them plan, build, test, and review the code in isolation. The key twist is that the agents live inside the sandbox, not outside it, so they can't mess up your real environment or steal your main credentials. You just mount the box, give it a task, and watch the trace stream back to your machine.

📰 https://news.ycombinator.com/item?id=49412872

🆔 @hackernewsgithubprojects

📦 url-kaist/multi-mapcher

Multi-Mapcher: Mapping Without Loop Closures

Multi-Mapcher is the mapping tool that aligns different LiDAR sensors without needing to find loop closures. For autonomous vehicles, building a single map from multiple sensors usually requires detecting when the robot returns to a previous spot, a process that is often unstable. This project solves that by using a method that tolerates errors in the alignment. It allows you to merge maps from different sensors, like a standard radar and a long-range lidar, into one accurate picture of the environment. By removing the need for that difficult detection step, it makes the mapping process more reliable.

🆔 @hackernewsgithubprojects

📦 steipete/camsnap

Camsnap: The Command Line Camera Remote

Camsnap lets you turn your terminal into a remote control for your security cameras. Instead of digging through a messy web portal, you type a simple command to grab a still photo or record a short video clip from any connected device. It’s built for people who live in the command line, making it easy to script your own automations. The coolest part is the motion watch feature. You can tell the tool to keep an eye on a specific camera and automatically trigger any shell command you want whenever it detects movement.

🆔 @hackernewsgithubprojects

📦 x-glacier/kali-pentest

The AI Hacker That Needs Your Permission

An AI agent just learned to plan a full-scale network intrusion on its own. Kali Pentest is a skill pack that teaches coding assistants like Claude Code how to act like a professional security tester, giving them a brain with over two hundred specialized tools and fifteen detailed playbooks. Instead of just guessing, the agent maps out a strategy, picks the right tools for the job, and adapts as it goes. The coolest part is that it pauses to ask you for approval before doing anything risky.

🆔 @hackernewsgithubprojects

📦 dreamzero0/dreamzero

DreamZero: The Robot That Predicts the Future

DreamZero makes robots act like they can see the future, letting them perform complex tasks without ever seeing them before. It’s a world action model that simultaneously predicts what the video will look like and what moves the robot should make, essentially dreaming out the next step in real time. This is huge because it solves the massive data problem in robotics; instead of needing thousands of hours of practice for every new task, this system generalizes instantly to unseen challenges. It even topped major leaderboards by learning entirely from standard datasets, proving that predicting the visual world is the key to smart, flexible robot behavior.

🆔 @hackernewsgithubprojects

📦 inevolin/k8s-cpu-limits-analyzed

K8s CPU Limits: The Hidden Performance Killer

This analysis proves that setting CPU limits in Kubernetes is silently destroying your application performance. While most developers treat limits as safe headroom, this repository demonstrates that they actually cause severe throttling, making apps up to four times slower without raising any red flags on standard dashboards. The project provides a clear, evidence-based guide explaining why removing these limits while keeping CPU requests is the only way to fix unexpected slowness and hidden memory issues. The core takeaway is simple: stop setting CPU limits to avoid artificial bottlenecks, and instead rely on requests to guarantee your share of the processor.

📰 https://news.ycombinator.com/item?id=49296939

🆔 @hackernewsgithubprojects

📦 vvedantb/eva

EVA: Cloud AI Coding Agents That Work While You Sleep

EVA is the cloud agent orchestrator that lets AI coding assistants work on your code while you focus on your day job. It spins up safe, isolated cloud environments where AI agents can run long tasks, browse live previews, and write code without touching your local machine. You simply point it at a GitHub repo, drop in a task, and watch it go. When it’s done, you get a clean pull request with diffs to review—no local setup, no babysitting, just shipped work.

📰 https://news.ycombinator.com/item?id=48875320

🆔 @hackernewsgithubprojects

📦 lzyang2000/perceptive_cbf_rl

PAC-MAN: Humanoid Dodgeball Safety

A humanoid robot just dodged a thrown ball nineteen times out of twenty without falling, using only a simple depth camera on its head. This is PAC-MAN, a research project from Caltech that teaches a Unitree G1 robot to avoid physical harm. Instead of just learning to move, the robot is trained with strict safety rules that act like invisible barriers. It watches a small, sparse image of the ball approaching and instantly decides whether to duck, step left, or step right. The coolest part is that it works on the real robot right away, no extra tuning needed.

🆔 @hackernewsgithubprojects

📦 benja/ask

Ask: The Terminal AI That Keeps Its Hands Off

Ask is the terminal command that hands you the answer without taking control of your files. Most coding tools edit your code behind the scenes, but Ask runs in read-only mode, forcing you to type the commands yourself. You ask a question, get a quick reply, and your shell stays right where you left it. It supports major AI agents like Codex and Claude, saving every session so you can pick up right where you left off. This is the perfect way to learn from an AI without worrying about it accidentally breaking your project. Stop letting bots code for you and start understanding the tools in your own terminal.

📰 https://news.ycombinator.com/item?id=49412440

🆔 @hackernewsgithubprojects