📦 rudderlabs/devbox

DevBox: Your New AI Engineering Teammate

DevBox is the autonomous engineering agent that turns a simple ticket into a finished, reviewed pull request without you touching a key. You hand it a task from your project tracker, and it plans the change, writes the code, runs the tests, and critiques its own work before opening a draft for you to merge. The coolest part is how it keeps your secrets safe; it runs in isolated sandboxes with only dummy tokens, so your real credentials never sit inside the environment. It even learns from your codebase over time, making it sharper with every task.

🆔 @hackernewsgithubprojects

📦 mutonby/agentring

Your AI Agent Calls Your Phone

Stop waiting for email alerts to reach your coding assistant. With Agent Ring, the AI agent actually rings your phone like a standard phone call to notify you about critical tasks. The voice AI explains the situation, you speak your instructions, and the agent receives the transcript to continue its work. It runs entirely on your own server, so you never pay per-minute fees. This open-source tool bridges the gap between digital automation and real human interaction. It turns passive notifications into active conversations. Your phone remains locked, but you stay in the loop. No third-party phone numbers are needed.

📰 https://news.ycombinator.com/item?id=50279324

🆔 @hackernewsgithubprojects

📦 wenyuchiou/awesome-agentic-ai-zh

Awesome Agentic AI: The Only Map You Need to Build Real AI Agents

A single, carefully curated map exists that guides you from zero to building reliable AI agents without the usual overwhelm. The repository called awesome agentic ai zh solves the biggest problem in AI learning: knowing what to learn next. It provides over 240 hand-picked resources and hands-on exercises in a clear, step-by-step roadmap. Instead of drowning in random tutorials, you follow a structured path from basic language models to complex multi-agent systems. The coolest part is that it is written in three languages, making it accessible to a global audience. It cuts through the noise by focusing on practical skills you can actually use.

🆔 @hackernewsgithubprojects

📦 ghabix/spinecodex

SpineCodex: 10x Context for AI Coding

SpineCodex is the coding assistant framework that turns a single chat into a scalable tree of tasks. Instead of cramming every detail into one long conversation, it splits your work into independent branches. The coolest part is that it compiles finished sections into memory, giving your AI a ten times larger effective context window without losing focus. This means it solves complex coding problems faster and cheaper than standard tools, because it does not forget what it learned earlier in the session. If you want your AI to handle big projects without getting confused, this is the structure you have been missing.

🆔 @hackernewsgithubprojects

📦 thomasunise/casecite

CaseCite: Private AI for Legal Research

Bring your own private AI legal research workspace. CaseCite lets law firms upload their own case files and contracts to a system they fully control, rather than relying on public cloud services. The standout feature is how it handles citations. Instead of just giving you an answer, it pulls exact quotes from your uploaded documents and verifies them against real court records. This means you can trust that every claim in a draft or summary is backed by actual source text, not just generated guesswork. It’s a powerful way to use modern AI for legal work while keeping sensitive client data on your own servers.

📰 https://news.ycombinator.com/item?id=50273623

🆔 @hackernewsgithubprojects

📦 cisco-ai-defense/aibom

Cisco AIBOM: The AI Inventory Tool

Cisco AIBOM reads your actual source code to build a complete inventory of every AI asset hiding in your software, from models to hidden secrets. It goes beyond simple dependency checks by using an artificial intelligence agent to verify findings and eliminate false positives, ensuring you know exactly what is running in your containers. This tool solves the problem of unknown AI components by scanning seven programming languages to map out datasets, prompts, and guardrails. It transforms vague code into a clear bill of materials, helping teams track compliance and security risks. You finally get a structured map of your AI stack. No more guessing what your software is doing.

🆔 @hackernewsgithubprojects

📦 fazmin/cve-intel

CVE Intel: Fix the Right Bugs

CVE Intel is the local security tool that cuts through the noise of thousands of new vulnerability reports to tell you exactly what to patch first. Every day, hundreds of new flaws appear in public databases, most of which never actually get attacked, leaving developers overwhelmed and guessing which ones matter. This desktop app, written in high-speed Rust, solves that chaos by pulling data from forty different public sources and combining them into a single, simple risk score from zero to one hundred.

📰 https://news.ycombinator.com/item?id=50293260

🆔 @hackernewsgithubprojects

📦 cisco-ai-defense/model-provenance-kit

Model Provenance Kit: Know Your AI's Roots

Cisco’s Model Provenance Kit gives every open-source language model a verifiable identity card. Instead of trusting a filename, this toolkit analyzes the internal structure of a model’s weights to match it against a catalog of known architectures. It answers the critical question of whether the file on your disk is actually what its label claims to be. This prevents supply-chain attacks where malicious code could hide inside a disguised model. By providing a clear audit trail for AI assets, it turns vague trust into concrete verification. Now you can know exactly who built the intelligence you are running.

🆔 @hackernewsgithubprojects

📦 chen-985211/cleancode

CleanCode: The Command Center for Your AI Coding Crew

CleanCode turns your terminal into a visual command center where you can actually see your AI coding assistants working. Instead of juggling separate windows for tools like Claude or Codex, you drag them onto a shared canvas, run multiple agents in parallel, and watch them tackle different parts of your project at once. It’s like having a whiteboard where you can assign tasks to your digital workers and observe their progress in real time. You can even save these setups as reusable templates, so you don’t have to rebuild your workflow every time you start a new project.

📰 https://news.ycombinator.com/item?id=50265699

🆔 @hackernewsgithubprojects

📦 xninety9/printr

Printr: The Coziest Receipt Printer Ever

Print tiny paper tickets for your family, right from your Raspberry Pi. Printr turns a standard thermal receipt printer into a daily companion that prints your morning weather, a fresh sudoku, or even a little note with a photo attached. The best part? It can print a riddle with the answer hidden upside down so you have to flip the paper to check it. It runs entirely on a tiny computer and talks directly to the printer, meaning no cloud services required for your morning puzzle. If you want a bit of physical charm in your smart home, this is the project to try.

🆔 @hackernewsgithubprojects

📦 bryceyuuu/run

Run: The Ultimate Global Relocation Map

Run is the relocation handbook that maps every legal way to move abroad from a single passport. It cuts through the noise of outdated blogs by tracking one hundred and ten specific visa routes across thirty-six European nations and thirteen global hubs. The standout feature is its precision on changing rules, like Spain shutting down its golden visa or Japan suddenly tripling the money needed for a business permit. It is not a law firm, but a living map of exactly which doors are still open right now.

🆔 @hackernewsgithubprojects

📦 theconsensuslabs/noalloc

noalloc: The Zero-Allocation Cop for Go

A single line of code can quietly wreck your Go server's performance by secretly dumping data onto the heap. The project called noalloc acts as a strict guardrail that scans your entire codebase to catch every hidden memory leak before it ships. It works like a strict editor that flags any function that steals memory from the garbage collector, forcing you to fix the problem instead of ignoring it. This tool is interesting because it shifts the burden from manual auditing to automated enforcement, giving developers a clear, line-by-line map of where memory actually goes. It also provides pre-built containers that never expand, making high-speed coding safer.

📰 https://news.ycombinator.com/item?id=50278645

🆔 @hackernewsgithubprojects