🛠️ Hacker News Tools

[Other] Show HN: Pathetic – Java pathfinding: 10k concurrent paths in ~7ms It started with drones. My co-founder and I were building a Minecraft plugin where drones had to navigate through unloaded chunks - terrain that doesn&#x27;t even exist in memory yet. No existing library had any concept of it, which led to drones being stuck in non-existence.<p>Sloppy at first, tightly coupled to Minecraft&#x27;s API, we built our own. But somewhere along the way we realized: the problem wasn&#x27;t Minecraft. It was how Java pathfinding is built in general - object-heavy, GC-hostile, and single-threaded by assumption.<p>So I split the project. The core became its own thing: pure Java 8+ for compatibility, featuring a zero-allocation primitive heap instead of a theoretically optimal FibonacciHeap, because pointer chasing and cache misses were unacceptable. 100% async, even in a single-threaded, async-hating Minecraft hell, and a processor pipeline providing full extensibility. No game engine, yet the possibility to adapt to your environment.<p>Sometimes it isn&#x27;t the algorithm that needs optimization, but the environment it lives in. Pathetic is full of these micro-optimizations: a BloomFilter lookup before the expensive closed set check, a multi-metric squared heuristic to kill Math.sqrt() overhead, and bit-packing 3D coordinates into a single primitive long. In order to get Java fast, you need to treat it like C++.<p>The result: 10k concurrent paths in ~7ms, where most libraries quietly die. A 20k distance path (equivalent to 100k+ nodes) in ~60ms without bloating the heap. The README is... a choice. Judge it, because it does the same to you.

[Other] Show HN: I built an API for agents visiting my personal website

[CLI Tool] Show HN: Codeberg (Forgejo) CLI, built with Xclif

[Other] Agents for financial services and insurance

[Other] Show HN: Octopus Code Review is now free for OSI-licensed repos

[Other] A polynomial autoencoder beats PCA on transformer embeddings

[Other] Show HN: Retroguard – Verifiably secure AI guardrails Cryptographically secure and verifiably robust protection with drop-in integration and outcome-based pricing

[Other] CVE-2026-31431: Copy Fail vs. rootless containers

[CLI Tool] Show HN: Kanban-CLI – a web UI for local Markdown todo lists As we all are, I&#x27;ve been experimenting with ways to reduce external saas spend, and continually bring traditionally external pieces of context (prs, docs, trello boards) into the one mono repo.<p>I have toyed with a markdown todo list and setting claude&#x2F;codex into a &quot;ralph&quot; loop previously, but always missed the web ui for organising, planning, and tracking things in progress.<p>So codex and I build kanban-cli, an installable or npx-runnable cli util that takes one or more markdown files, and uses markdown headers and todo syntax to track and manage agentic tasks.<p>Let me know your thoughts, or if you&#x27;ve got a more convenient way of managing tasks!<p>I was tempted to link to the project landing page <a href="https:&#x2F;&#x2F;kanban-cli.vochsel.com" rel="nofollow">https:&#x2F;&#x2F;kanban-cli.vochsel.com</a> as I made a cute example demo, but linking to the repo felt less self-promo-y.<p>Cheers!

[Other] Show HN: I indexed 8,643 BSides talks across 227 chapters and 6 continents Hi HN,<p>I&#x27;m Roland, and for the past few weeks, I&#x27;ve been building AllBSides — a directory of every BSides conference talk uploaded to YouTube. As of today, 8,643 talks from 5,927 speakers across 227 chapters in 68 countries. Combined runtime is 280 days. The transcripts come to about 60 million words.<p>The archive came together in stages:<p>1. Manually map every BSides chapter&#x27;s YouTube channel 2. Pull every video and transcript from Supabase 3. Run each transcript through Haiku for tag extraction (tools, topics, difficulty, team, talk style, research method, and much more) 4. Run results through Sonnet for categorization and dedup 5. Final pass goes through Opus for verification 6. Do a manual verification - at one time, the pipeline showed over 16k AI suggestions for manual verification. Today, most are resolved.<p>Total LLM cost so far: about €200. The whole pipeline is rebuildable from scratch.<p>Each talk gets its own page with embedded video, full transcript, speakers, tags, and &quot;related talks.&quot; Each tool&#x2F;framework&#x2F;protocol&#x2F;standard mentioned across the corpus gets its own page (3,968 distinct technologies tracked).<p>Some interesting facts I gathered while building it:<p>-(A) The site is currently 94% bot traffic. Of that, about 80,000 hits&#x2F;month are AI training crawlers (ClaudeBot, GPTBot, meta-externalagent). Within 7 days of the talks archive going live, all major AI labs had ingested the entire corpus. The discovery cascade was startling to watch in real time.<p>-(B) The taxonomy work was the hardest part. Distinguishing &quot;tools&quot; from &quot;frameworks&quot; from &quot;protocols&quot; from &quot;concepts&quot; sounds easy until you have 5,000 ambiguous extracted entities. The 3-tier LLM pipeline helped a lot — Haiku alone was too noisy, Opus alone was too expensive.<p>-(C) Top tools mentioned: Wireshark (343), PowerShell (342), Metasploit (332), Burp Suite (322), GitHub (296), VirusTotal (273), Docker (253), Splunk (251), Nmap (247), MITRE ATT&amp;CK (237). The list reflects what BSides talks actually discuss, not what vendors curate.<p>-(D) May is the peak BSides month — 29 events, 17% of all events with dates.<p>-(E) The top 1% of talks (86 videos by view count) account for 51% of all viewership. The other 99% are deeply niche, often the only video record of a specific technique.<p>The stack is intentionally lean: Go, SQLite, vanilla JavaScript, BunnyCDN. Static rendering at build time. No frameworks, no client-side state. The site costs about €50&#x2F;month to run.<p>The data behind this post and much more can be found in the site footer, under the link &quot;stats&quot;.<p>Happy to answer questions about the data pipeline, the taxonomy decisions, or what the AI crawler patterns looked like as the archive went live. Feedback on what to build next is genuinely welcome — I&#x27;m a solo dev figuring this out as I go.<p>— Roland (parkado)

[Other] Formatting a 25M-line codebase overnight

[Other] Show HN: Agent-evals – Claude skill to build your own evals I’ve spent the past 10 years working on AI in finance, with much of that time focused on building evaluation systems for production environments.<p>As agents become more widely adopted, more software engineering and product people have start building them. But I’ve noticed that many teams are not yet fluent in systematic evaluation, or in the processes needed to keep agent quality high over time.<p>For large organizations, that gap is rarely the bottleneck due to dedicated teams. But after speaking with a number of startups, it became clear that building strong, up-to-date evals is much harder in a fast startup, especially when the team does not have a data science background.<p>So I tried to condense as much of my experience as possible into a Claude Skill: a practical starting point for evaluating your agent.<p>The idea is simple: tell Claude you need evals, and it will set up a solid baseline directly in your codebase - that&#x27;s it! The evals will follow patterns I&#x27;ve seen many times before, and will get you a summary of what your agent does well and what it doesnt.<p>Looking forward to your feedback!

[Other] Google is discontinuing its free web search index for developers

[Other] White House Considers Vetting A.I. Models Before They Are Released

[Other] Days without GitHub incidents

[Other] Show HN: Muesli – If Granola and Wisprflow had an open source on device baby Hey folks, I am the developer behind muesli - which is your one stop app for all your speech to text needs, be it voice dictation or meeting transcriptions that runs on device on your Apple Neural Engine using CoreML based STT models (Parakeet, Whisper, Cohere transcribe). Everything is open source and we are at 160 stars - au naturale - would love for folks to use it and contribute further to the development

[Other] Show HN: Bonsai 1.7B ternary model at 442T/s on M4 Max We took a recently released Bonsai 1.7B ternary model from PrismML (<a href="https:&#x2F;&#x2F;github.com&#x2F;PrismML-Eng&#x2F;Bonsai-demo" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;PrismML-Eng&#x2F;Bonsai-demo</a>) and ran our agentic evolution search on it for 6 hours to optimize the Metal kernels. The search was fully autonomous.<p>Measured against unmodified upstream llama.cpp at the same Bonsai&#x2F;Q2_0 commit, same M4 Max:<p>- tg128: 309.82 → 442.42 t&#x2F;s (+42.0%)<p>- pp512: 4250.32 → 4622.63 t&#x2F;s (+8.8%)

[Other] Show HN: Replacing spec-driven development with just facts I had a lot of issues with spec-driven approaches, agents are too readily producing fluff, large projects have so many specs agents start making mistakes maintaining them. There&#x27;s a constant consistency tax.<p>In the end every spec is just a bunch of facts, so I decided to leave that and throw away everything else while making it friendlier for agentic use.<p>Introducing facts - skills and CLI for agents to use facts-driven development. <a href="https:&#x2F;&#x2F;github.com&#x2F;av&#x2F;facts" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;av&#x2F;facts</a>

[Other] PyInfra 3.8.0 Is Out

[Monitoring/Observability] Setting up server monitoring for a Rails app on Hatchbox