Feed aggregator

OP here. We built Tharos because we were tired of "security theatre" tools that produce read-only reports.

Tharos is an open-source security engine written in Go. Key features:

Polyglot: Scans Go, Python, JS/TS (AST-based + Regex). Interactive: tharos --interactive opens a TUI to fix issues one-by-one. CI-Ready: Returns strict exit codes for blocking builds. Visual: HTML reports and rich terminal UI. Repo: https://github.com/chinonsochikelue/tharos Docs: https://tharos.vercel.app

Would love feedback on the rule engine and the TUI ux!

Comments URL: https://news.ycombinator.com/item?id=46927353

Points: 1

# Comments: 0

Article URL: https://simonsafar.com/2024/win32_lights/

Comments URL: https://news.ycombinator.com/item?id=46927347

Points: 1

# Comments: 0

Article URL: https://www.ibli.com/IBLI%20OnePagers%20The%20Plays%20Summarized.pdf

Comments URL: https://news.ycombinator.com/item?id=46927343

Points: 1

# Comments: 1

Forget dim solar-powered lights: Here's what actually happens when you upgrade your patio with inexpensive, waterproof, smart LEDs.&

Article URL: https://pydantic.dev/articles/the-logs-i-never-read

Comments URL: https://news.ycombinator.com/item?id=46926829

Points: 1

# Comments: 0

Article URL: https://idratherbewriting.com/blog/bakhtin-collapse-ai-expressive-writing

Comments URL: https://news.ycombinator.com/item?id=46926812

Points: 1

# Comments: 0

Hi HN,

I built a real-time UART analyzer using an ESP32-S3 and a Windows GUI.

It captures high-speed UART data and shows live statistics and visualization on PC.

Main features: - Up to 921600 baud capture - Real-time throughput monitoring - RAW hex viewer - Frame decoding - Health / buffer status - Standalone Windows EXE

How it works: Target MCU TX → ESP32-S3 RX → USB → PC Analyzer

You just connect your device’s TX pin to the ESP32-S3 RX pin and run the PC app.

Supported devices: STM32, ESP32/ESP8266, NRF52, RP2040, Arduino, and other UART devices.

GitHub: https://github.com/choihimchan/linkscope-bpu-uart-analyzer

Demo GIF is in the README.

I’m actively improving this project and would really appreciate feedback, ideas, and bug reports.

Thanks!

Comments URL: https://news.ycombinator.com/item?id=46926807

Points: 1

# Comments: 0

Article URL: https://github.com/user19870/cppsp

Comments URL: https://news.ycombinator.com/item?id=46926788

Points: 1

# Comments: 1

Article URL: https://www.washingtonpost.com/health/2026/01/24/fractyl-glp1-gene-therapy/

Comments URL: https://news.ycombinator.com/item?id=46926746

Points: 1

# Comments: 1

Article URL: https://spectrum.ieee.org/wikipedia-at-25

Comments URL: https://news.ycombinator.com/item?id=46926716

Points: 1

# Comments: 3

Article URL: https://reviewreact.com

Comments URL: https://news.ycombinator.com/item?id=46926713

Points: 2

# Comments: 1

Article URL: https://zenodo.org/records/18514533

Comments URL: https://news.ycombinator.com/item?id=46926695

Points: 1

# Comments: 0

My experience using Claude Code is that I will detail work I want it to do, then it does it, then I spend quite a bit of time with it trying to figure out why it didn't work. Putting in debugging sometimes feels like half of the code it is writing. At the same time, I'm using more tokens by default with this "better" version of Claude. Of course I don't think this is 'on purpose' but it feels like there should be some discount on tokens being used to fix the bugs being caused by the bugs being introduced. Curious what others' experience are and if you've developed any strategies to address this.

Comments URL: https://news.ycombinator.com/item?id=46926656

Points: 1

# Comments: 0

Hi HN,

I built a CLI that unifies configuration for AI coding assistants.

Problem: Modern AI-powered development uses multiple tools (Cursor, Claude Desktop, GitHub Copilot, various CLI tools). Each has its own config file for MCP (Model Context Protocol) servers. Adding a server means manually updating 4-6 different configs.

Solution: agents gives you one source of truth. Define MCP servers once, sync everywhere.

Example: .agents/ ├── agents.json → Your config (committed) ├── local.json → Secrets (gitignored) $ agents sync → Updates all tool configs

Key decisions: - Convention over configuration (sensible defaults) - Security-first (secrets separated, auto-gitignored) - Interactive setup wizard for easy onboarding - 70+ tests (Vitest)

Supports: Cursor, Claude Code, Gemini CLI, Codex, Copilot, Antigravity

GitHub: https://github.com/amtiYo/agents npm: @agents-dev/cli

v0.7.7 released today. MIT licensed.

Looking for feedback on: 1. Is this solving a real problem? 2. What other tools should it support? 3. Thoughts on the architecture?

Happy to answer questions about the implementation or design decisions.

Comments URL: https://news.ycombinator.com/item?id=46926648

Points: 1

# Comments: 0

People

Comments URL: https://news.ycombinator.com/item?id=46926638

Points: 1

# Comments: 0

Article URL: https://twitter.com/JJackBrandt/status/2019852423980875794

Comments URL: https://news.ycombinator.com/item?id=46926617

Points: 2

# Comments: 0

Hi HN,

This is the first public launch of Writtte, an open-source, drafting-first editor built for writing and publishing articles.

We built this after noticing that most writers use 4-5 different tools just to draft and publish a single article. The biggest pain was not writing itself, but reformatting content every time it moved to Medium, Substack, WordPress, or a Markdown-based blog.

Writtte focuses on one thing: article drafting without format breakage.

What it does:

1. A psychology-designed editor UI that separates drafting from editing

2. AI-assisted + algorithm based grammar, clarity, tone, and refactoring (triggered manually)

3. One click copyto Medium, Substack, WordPress (Gutenberg), and Markdown

4. When you paste content into those platforms, formatting stays the same

5. Public draft sharing so others can review before publishing

This is completely open source: https://github.com/writtte/writtte

Public roadmap: https://github.com/orgs/writtte/projects/1

What we are working on next:

1. LaTeX support

2. Mermaid diagrams

We are building this in the open and would really appreciate feedback, especially from people who write articles or long-form content regularly.

Thanks for reading.

Comments URL: https://news.ycombinator.com/item?id=46926606

Points: 1

# Comments: 0

Article URL: https://www.youtube.com/watch?v=e9FUdOfp8ME

Comments URL: https://news.ycombinator.com/item?id=46926597

Points: 1

# Comments: 0

Article URL: https://www.hpcwire.com/off-the-wire/brookhaven-labs-rhic-concludes-25-year-run-with-final-collisions/

Comments URL: https://news.ycombinator.com/item?id=46926576

Points: 2

# Comments: 0