📦 sgkdev/packet_edit_meme

packet_edit_meme

packet_edit_meme is the security proof-of-concept that demonstrates how a subtle Linux kernel bug can let an unprivileged user silently hijack a system. By exploiting a flaw in how the kernel handles network packet modifications, the tool targets the system's page cache, which stores files in memory for fast access. It tricks the cache into overwriting a critical system file with custom code, granting the user full administrative access. It shows exactly how a minor memory-handling mistake can compromise overall system security, highlighting a critical patch that developers and administrators need to apply to keep their systems secure.

📰 https://news.ycombinator.com/item?id=48688054

🆔 @hackernewsgithubprojects

📦 axtonliu/axton-obsidian-visual-skills

axton-obsidian-visual-skills

You can now design beautiful, hand-drawn Excalidraw sketches, structured Mermaid flowcharts, and color-coded Obsidian Canvas mind maps directly from your terminal. This is possible because axton-obsidian-visual-skills turns these complex visual formats into simple, text-based extensions for your command-line AI assistant. Instead of messing with complex setup steps, you just install these markdown-based instructions to let your assistant generate valid, beautifully structured diagram files on demand. It solves the headache of manual positioning and broken syntax by calculating node spacing automatically, preventing code errors, and using charming handwriting fonts. Grab this setup to instantly turn plain text ideas into gorgeous, ready-to-use visuals.

🆔 @hackernewsgithubprojects

📦 off-grid-ai/off-grid-ai-mobile

off-grid-ai-mobile

Run massive artificial intelligence models entirely on your phone with zero internet connection. The off-grid-ai-mobile project is a complete on-device suite that lets you download and run large language models, transcribe voice with Whisper, and even generate images using Stable Diffusion natively on your hardware. It completely eliminates privacy worries because absolutely no data ever leaves your device. Plus, it lets you upload PDFs and text files to search through a local database using your phone's processor. It is basically like having a personal supercomputer in your pocket that works even in the middle of the woods.

🆔 @hackernewsgithubprojects

📦 maymhappy/meshripple

MeshRipple

AI generators often struggle to create clean 3D models because they lose track of the big picture, leaving behind messy holes and disconnected pieces. MeshRipple solves this by growing 3D meshes outward from an active border, much like ripples expanding across the surface of a pond. By combining a search strategy that aligns generation with the model's actual physical shape and a smart memory system that remembers long-range connections, it keeps the entire geometry physically complete. If you want to see how smart algorithms are finally mastering clean, artist-grade 3D model generation, you should definitely check this out.

🆔 @hackernewsgithubprojects

📦 clawdbrunner/captcha-solver

captcha-solver

captcha-solver is the browser automation toolkit that actually stops security challenges before they even load. Instead of just paying for third-party solve APIs right away, it uses a clever three-layer defense system starting with a heavily customized browser modified with over fifty low-level patches. This customized browser mimics real human behavior so perfectly that major websites let it pass without ever showing a security puzzle. If a challenge does pop up, the system automatically tries free interactive clicking tools before falling back to premium paid solvers. It is a brilliant, cost-saving approach to web scraping that keeps your automated tasks running smoothly.

🆔 @hackernewsgithubprojects

📦 slackapi/slack-mcp-plugin

slack-mcp-plugin

You can now search, read, and post to Slack directly from your AI developer tools. The slack-mcp-plugin brings Slack's entire workspace functionality straight into Claude Code and Cursor. Instead of jumping back and forth to copy-paste updates, you can run natural commands to summarize active channels, gather team standups, search across public threads, or even draft and queue announcements. For developers, it makes building Slack apps much easier by providing handy tools to generate and validate Block Kit layouts on the fly. It is a brilliant way to keep your workflow unified without leaving your editor terminal.

🆔 @hackernewsgithubprojects

📦 momenbasel/puremac

PureMac

Most Mac cleaners ask for deep disk access only to push paid subscriptions, track your usage, and shout fake warnings about your computer being in danger. The open-source app puremac does the exact opposite by strictly moving deleted items to the trash instead of permanently unlinking them. It uses a clever ten-level matching system to sniff out hidden preference files, launch agents, and orphaned caches left behind by deleted applications. There are no tracking scripts, no premium upsells, and no dramatized storage warnings. It gives you a clean, honest look at your storage and hands you full control over reclaiming your drive.

🆔 @hackernewsgithubprojects

📦 mlpc-ucsd/pixarmesh

PixARMesh

PixARMesh is the 3D scene reconstruction tool that directly outputs clean, native computer graphics meshes from a single camera snapshot. Instead of relying on messy intermediate steps like volumetric grids or complex mathematical functions that you have to decode later, it models entire rooms and objects as native geometric meshes. By predicting both where objects are and what they look like in a single, unified sequence, it simplifies the transition from a flat picture to a fully realized 3D environment. This makes it incredibly easy to turn real-world photos into usable assets for game engines, virtual reality, or physical modeling.

🆔 @hackernewsgithubprojects

📦 softcane/human-state-skills

human-state-skills

human-state-skills is the AI assistant toolkit that changes its entire communication style based on your current mental energy. Most AI tools give you a giant list of steps when you ask for advice, but if you are exhausted or overwhelmed, a wall of text is the last thing you need. This project lets you activate specialized modes like overloaded, foggy, or burnout directly in your AI chat. Instead of telling you to re-plan your whole week, the AI might give you just one single action to take next, like putting a document on the table, and tell you exactly where to stop. It adapts the assistant to match how your...

🆔 @hackernewsgithubprojects

📦 arya51-ai/obsidian-vault-sync

obsidian-vault-sync

obsidian-vault-sync is the local script that turns your messy Claude conversations into a beautifully organized knowledge graph inside Obsidian. Instead of letting your best AI chats vanish into browser history, this tool automatically pulls your Claude sessions, formats them into clean markdown, and files them away by project. The coolest part is how it handles memory. It actually symlinks Claude's own memory files directly into your vault, meaning the notes Claude keeps about your work become real Obsidian notes you can link and search. It even updates notes as your conversations continue, keeping your personal knowledge base perfectly in sync.

🆔 @hackernewsgithubprojects

📦 freecomputinglab/rheo

Rheo

Rheo is a publishing engine that compiles a single folder of Typst documents into a PDF document, a static HTML website, and an EPUB e-book all at the exact same time. Instead of managing separate source files or fighting with fragile conversion tools, you write your content once using clean Typst markup and let this tool handle the rest. It features a built-in development server with automatic browser refresh for real-time editing, and it even resolves links between documents automatically across every format. It is a brilliant way to publish beautiful books, websites, and documents simultaneously with zero extra effort.

📰 https://news.ycombinator.com/item?id=48730169

🆔 @hackernewsgithubprojects

📦 pzqpzq/lsf_mdia

lsf_mdia

Compress complex language model reasoning into ultra-compact symbolic dialects instead of wasting time and money on long, rambling chain of thought explanations. By letting different AI models invent, refine, and route their own custom symbolic shorthand, this system cuts down token usage dramatically while actually boosting accuracy on complex tasks. It treats communication as a dynamic social process where models create task-specific codes tailored to each other. This elegant approach proves that machines do not need to think in human paragraphs to solve tough problems. Check out the repository to see how language models can design their own efficient languages.

🆔 @hackernewsgithubprojects

📦 frumu-ai/tandem

Tandem

Tandem is the desktop safety manager that finally gives you total control over what autonomous AI agents can actually do on your machine. Instead of letting AI tools run wild with your files, memory, and code, this app steps in as an active gatekeeper to monitor and approve every single action they take. You get real-time approvals, clean audit trails, and pre-configured skill packs for specialized tasks like data analysis or legal research. It acts like an operating system sandbox with human-in-the-loop oversight. Keep your local files safe while letting your AI agents do the heavy lifting.

🆔 @hackernewsgithubprojects

📦 internscience/agents-a1

agents-a1

A relatively compact thirty-five-billion parameter AI model can match the performance of trillion-parameter giants simply by scaling up its reasoning horizon. The agents-a1 project introduces an open-source mixture-of-experts model trained on massive agentic trajectories averaging forty-five thousand tokens. By connecting external knowledge, actions, and feedback, it tackles incredibly complex, multi-step tasks across six domains like scientific research and coding. It uses a clever multi-teacher training approach to pack specialized expert knowledge into a single deployable model. It is a brilliant reminder that smarter training and longer reasoning paths can easily beat sheer model size.

🆔 @hackernewsgithubprojects

📦 epoch-research/mirrorcode

MirrorCode

An artificial intelligence model once spent nineteen straight days working on a single problem in this evaluation suite. This is mirrorcode, a rigorous benchmark designed to test the limits of machine learning on complex programming tasks. Built on the Inspect framework, it forces artificial intelligence to tackle real-world software engineering challenges using actual source code and documentation from popular tools like Ruff and JQ. Instead of simple coding puzzles, it evaluates how systems handle deep, multi-step engineering workloads over long periods. It is a fascinating look at how researchers measure the boundary between human developers and next-generation software agents.

🆔 @hackernewsgithubprojects

📦 fansunqi/videotool

VideoTool

Large multimodal AI models can now analyze video questions with unprecedented accuracy by strategically combining specialized temporal and spatial tools. videotool is an open-source framework from NeurIPS 2025 that equips language models with a collaborative toolkit containing frame selectors, object trackers, and image captioners. Instead of relying on a single AI to guess its way through a video, this project coordinates these specialized tools to progressively narrow down and isolate exactly where and when an action occurs. It dramatically reduces errors and boosts video understanding performance, making it a must-watch helper for building smarter video applications.

🆔 @hackernewsgithubprojects

📦 open-gsd/gsd-core

Gsd Core

The software development companion that actually gets your code out the door without the typical chaos, gsd-core is a framework designed to turn manual terminal work into structured, issue-driven developer workflows. Instead of jumping between endless AI tools, it uses specialized agents and built-in commands to coordinate everything from planning and code reviews to debugging and deployment. It acts like an automated project manager and quality assurance team living right inside your command line, verifying your code changes and updating your documentation dynamically as you work. Give it a spin on your next project to see how easy shipping code can really be.

🆔 @hackernewsgithubprojects

📦 danielmiessler/lifeos

lifeos

Build a personal digital command center that aligns your daily actions with your biggest life goals. This repository, called lifeos, provides the actual framework to run what it calls a personal AI infrastructure. Instead of juggling random productivity apps, it combines your goals, habits, and tasks into a single system powered by a local dashboard and a digital assistant. The core engine analyzes your current state, maps out where you want to go, and helps you take the next step. It is a fascinating way to use modern AI as a highly personalized coach that keeps you focused on what actually matters.

🆔 @hackernewsgithubprojects

📦 cloudflare/skills

Cloudflare Skills

AI coding assistants can now get direct, native instructions to build and manage full cloud applications on Cloudflare. The skills repository is a collection of pre-made capabilities designed for popular AI agents like Claude Code and Cursor. Instead of relying on outdated training data, this project equips your agent with specialized knowledge for databases, AI models, and secure code sandboxes. It even includes remote servers that let your AI browse live documentation and manage active deployments. It is a fantastic way to turn a generic coding assistant into a dedicated cloud engineer that gets things right the very first time.

🆔 @hackernewsgithubprojects

📦 mindmuxai/brain.md

brain.md

Keep your coding agents on the exact same page across every single session by giving them a shared, persistent memory. This zero-dependency command line tool scaffolds a plain markdown brain file directly inside your repository. Instead of losing critical decisions, project constraints, and architectural context the moment a chat ends, agents like Claude Code read and write to this local file to instantly recall why choices were made. By routing all memory updates through a simple command line interface, the project structurally prevents malformed edits, keeping your project's durable knowledge base clean, portable, and completely independent of any specific AI model.

📰 https://news.ycombinator.com/item?id=48717245

🆔 @hackernewsgithubprojects