🛠️ All DevTools

Do agents.md files help coding agents? <a href="https:&#x2F;&#x2F;xcancel.com&#x2F;rasbt&#x2F;status&#x2F;2063649136323252397" rel="nofollow">https:&#x2F;&#x2F;xcancel.com&#x2F;rasbt&#x2F;status&#x2F;2063649136323252397</a><p><a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2602.11988" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2602.11988</a>

Show HN: Web Speed – A shared web-map registry for AI agents (MCP, open source) Hey everyone, I&#x27;ve been working on a tool called Web Speed. It takes web pages full of HTML and parses them into an easily readable sitemap for an AI agent. By ensuring the AI doesn&#x27;t have to analyze the full HTML from the page or screenshots of the page, it makes the agent a lot cheaper and faster. It&#x27;s MCP-native, so any AI that supports MCP can run it and take control of your browser for you. There are both “normal” and post-auth versions live on GitHub. What I think is the coolest thing about the project is the global cache of sitemaps. When a user visits a site, it takes the sitemap and sends it up to the server so that later, a person can get that sitemap from the cache, further speeding up the agent. Right now, only the paid API version can access the cache. I would love your thoughts on the project, any ideas, criticisms, or concerns. I really just want to see if people think that this would be a useful tool. Please reach out if you have any questions about Web Speed. Repo (OSS): <a href="https:&#x2F;&#x2F;github.com&#x2F;Dominic-Pi-Sunyer&#x2F;web-speed-oss" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Dominic-Pi-Sunyer&#x2F;web-speed-oss</a> Repo (Post-auth): <a href="https:&#x2F;&#x2F;github.com&#x2F;Dominic-Pi-Sunyer" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Dominic-Pi-Sunyer</a> API and Docs: <a href="https:&#x2F;&#x2F;getwebspeed.io" rel="nofollow">https:&#x2F;&#x2F;getwebspeed.io</a><p>Thanks, Dominic

Show HN: I put my Claude Code rate-limit burndown in the status line

Tech sell-off widens as South Korea index plunges

DeepSeek V4 Pro beats GPT-5.5 Pro on precision

APC–2 – A professional record cutter for producing original playback discs

The Smallest Brain You Can Build: A Perceptron in Python

Firefox Merges Support for Vulkan Video Decoding

Show HN: Nightwatch, The open-source, read-only AI SRE nightwatch is a local-first, read-only layer on top of your monitoring. it groups alert storm into incidents, flags noisy checks and has an agent that can investigate for you live systems. You can e.g. jump from the incident into the agent directly.<p>the reason for this weekend project is that we had a kubernetes upgrade that went wrong, and at some point a rollback wasn&#x27;t possible anymore, so it had to be fixed live during the night while several problems came together. We run a lot of different systems, on-prem and several Kubernetes clusters, and in a situation like that you spend most of the time just figuring out what is actually broken and where.<p>So i thought that it would be pretty cool to have eyes in the dark in each system that can talk to your &quot;brain&quot;.<p>so the idea is to put a baby owl into each environment. Each owl runs where the systems live, keeps that environment&#x27;s credentials local, and only dials outbound to a central brain, so there is no inbound hole into prod. It exposes a set of read-only skills, and the agent uses them to gather evidence and form a root-cause hypothesis, so the on-call engineer starts with a head start instead of from zero.<p>read-only for now, i don&#x27;t trust it near prod yet and honestly neither should you.<p>llocal-first for easy self-hosting and to keep credentials on your side. the clustering and recommendations run fully offline with no llm at all. the agent needs a tool-calling llm, you can point it at a remote one, or self-host one (ollama etc.) if you want to stay fully offline.<p>for non selfhosters: before every remote llm call, nightwatch strips real secrets (unrestorable) and swaps identifiers like ips, hostnames and paths for reversible placeholders, so the model only sees masked data while real values are restored only in the proposed commands and tool calls<p>Would love if you try it in your Systems

Thi.ng – open-source building blocks for computational design and art

How do you design a $30k electric pickup? Inside Ford's skunkworks

Leiden Declaration on Artificial Intelligence and Mathematics

Building from Zero After Addiction, Prison, and a Felony

An introduction to functional analysis for science and engineering

Show HN: Help SourceLibrary.org Translate the Renaissance SourceLibrary.org is currently the world’s largest online library of translated ancient sources. It contains over 15,000 books translated from dozens of languages (Latin, Greek, Chinese, Sanskrit, Arabic, Hebrew, Egyptian, etc)—from sacred texts to early science. Our vision is to make ancient wisdom accessible to both people and AI.<p>Did you know that less than 3% of the Latin Renaissance has ever been translated into English?<p>This is ironic, because translations of ancient Greek wisdom texts are historically credited with helping to ignite the Renaissance. Yet, the Renaissance itself was mostly written in NeoLatin and it is currently 97% untranslated.<p>For scale, the word count of SourceLibrary.org just exceeded English Wikipedia. In addition to the texts, there are hundreds of thousands of searchable images.<p>SourceLibrary is free and open-source (AGPL3). Scans are mostly public domain and translated texts are Creative Commons share-alike. All is accessible via MCP or API. Point Claude Code or Codex at sourcelibrary and see what it is like when it can search and quote source texts. E.g., ask Claude Code to compare alchemical practices in Latin, Chinese, Sanskrit etc — and ask for images.<p>SourceLibrary.org is an open source initiative of the Embassy of the Free Mind (<a href="https:&#x2F;&#x2F;www.EmbassyoftheFreeMind.com" rel="nofollow">https:&#x2F;&#x2F;www.EmbassyoftheFreeMind.com</a>), a Dutch nonprofit with 501c3 status and home to the <i>Biblioteca Philosophica Hermetica,</i> a UNESCO recognized “Memory of the World” rare book library. SourceLibrary is entirely supported by tax-deductable donations. Corporate sponsorships and partnerships are also welcome: <a href="https:&#x2F;&#x2F;sourcelibrary.org&#x2F;support" rel="nofollow">https:&#x2F;&#x2F;sourcelibrary.org&#x2F;support</a><p>We invite you to participate in this historical effort to translate the Renaissance and beyond. In addition to the digital impact, we hope to manifest long-term institutional support for the (global) human stewardship of these materials.<p>HN, please sign up for free and leave feedback (found at the bottom of every page). If something is annoying or improper or hard to use — please share. I read feedback almost every day. Maybe test out the librarian <a href="https:&#x2F;&#x2F;sourcelibrary.org&#x2F;librarian" rel="nofollow">https:&#x2F;&#x2F;sourcelibrary.org&#x2F;librarian</a> — it’s a research agent that can retrieve quotes and images in chat.<p>If this project resonates with you, please let me know what you find!<p>Derek<p>Acknowledgments: Thank you to my friends at playpowerlabs.com, getsmartpaper.com and Wisdom-Frontier.org for concept and development support and to <a href="https:&#x2F;&#x2F;www.frond.studio&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.frond.studio&#x2F;</a> for design support

Show HN: GentleOS – A pair of hobby OSes for vintage 32-bit and 16-bit PCs Hello HN,<p>I&#x27;ve been working on a simple OS for tinkering and running bare metal apps on vintage PCs.<p>Since I couldn&#x27;t quite decide whether to target pure 16-bit, or slightly more capable 32-bit machines, I ended up with two separate versions:<p>- GentleOS&#x2F;32 (<a href="https:&#x2F;&#x2F;github.com&#x2F;luke8086&#x2F;gentleos32" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;luke8086&#x2F;gentleos32</a>) works on i386+, requires 4MB of RAM and VGA display supporting 640x480x16 mode or any 256-color VESA mode.<p>- GentleOS&#x2F;16 (<a href="https:&#x2F;&#x2F;github.com&#x2F;luke8086&#x2F;gentleos" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;luke8086&#x2F;gentleos</a>) works on 80186+, requires less than 192KB of RAM and a CGA display supporting 320x200x4 mode.<p>You can find more details in the repos.

Anthropic, please ship an official Claude Desktop for Linux

Show HN: I made a better zsh autosuggestion tool that predicts your next command Hi everyone I just created Deja, a tool that instead of only surfacing commands that start with what you&#x27;ve typed, suggest what you actually want to run. I built it because I was using zsh autosuggestions but got tired of typing the same commands again and again. So Deja predicts your next command from your history.<p>Let me know what you think

Show HN: A virtual thermal printer for testing ESC/POS receipts

Misguided Misstatements Continue to Dismantle Biomedical Research in the U.S.