🛠️ All DevTools

[Other] Tracing HTTP Requests with Go's net/HTTP/httptrace

[Other] A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.

[Other] Show HN: Ktx – Open-source executable context layer for data agents Hi HN, we’re open-sourcing ktx. It’s an executable context layer that makes agents reliable on your data stack.<p>We built it after going through the experience of building production-grade data agents for dozens of companies. If you’ve also tried building them, or simply tried using Claude Code or Codex on your data warehouse, you’ll know that accuracy is the #1 issue. Agents are great at generating valid SQL, but it’s not always correct SQL.<p>To cite a few examples of “agents gone wrong”:<p>- Stale column + hidden business rule: when preparing a board report, a finance analyst asks Claude Code for “ARR by customer segment”, it derives ARR from multiple tables (subscriptions, plans, accounts), then groups by accounts.industry. But CC doesn’t know that this industry column was deprecated a few months prior, or that past board reports excluded paused subscriptions from the ARR calculation<p>- Join fanout: a data analyst at a retailer uses their company’s internal agent to prep a product revenue deck for a QBR. The agent joins orders to order_items, then sums orders.total_amount_cents grouped by order_items.product_id. The SQL runs fine, but each order’s revenue is repeated once per line item, which most people will miss if most orders only have 1 item<p>- Missing attribution logic: a marketing analyst asks Codex “Which campaigns drove the most revenue?” Codex joins marketing_touches to users to orders and groups by utm_campaign. But since each order can have multiple touches before purchase, the same order can be credited to first touch, last touch, every touch, or every campaign the user clicked before buying. If the agent chooses the method that doesn’t match the team’s attribution logic, they’ll make suboptimal decisions<p>To solve this at first we gave the agent more context through skills + a wiki-style knowledge base. That gives it some useful extra context but still relies on it writing the SQL without incorrect assumptions.<p>The next solution we explored was implementing a classic semantic layer. That solves the executable part, but they’re such a pain to build and maintain since they were made for legacy BI tools. Plus as a standalone tool, they lack all the useful context from unstructured data sources like internal docs.<p>So we built ktx and split it into 2 parts:<p>1. Business context goes in Markdown wiki pages that are auto-ingested and auto-populated<p>2. Queryable definitions go into YAML files that define tables, row grain, joins, measures, dimensions, filters, and filter groups<p>That way, when an agent needs a metric, it asks ktx for a measure, dimensions, filters, and filter groups instead of writing the whole query itself. ktx’s planner chooses the join path, uses grain and relationship metadata, catches issues like join fanout and chasm joins, and compiles the warehouse SQL, while utilizing the extra unstructured knowledge it has access to.<p>ktx is Apache 2.0. It can ingest from most warehouses (BigQuery, Snowflake, Postgres &amp; others), modeling tools (dbt, MetricFlow, LookML), BI tools (Looker, Metabase), doc tools like Notion, and corrections from user interactions.<p>Install manually:<p>npm install -g @kaelio&#x2F;ktx<p>ktx setup<p>Or give this prompt to your agent:<p>Run npx skills add Kaelio&#x2F;ktx --skill ktx and use ktx skill to install and configure ktx<p>We’d especially like feedback from people who’ve tried using Claude Code, Codex, or building custom agents on analytics warehouses. Where did they fail? And what did you try to make the answers more reliable?

[Testing] Creusot helps you prove your Rust code is correct

[Monitoring/Observability] Show HN: Multiplayer, a debugging agent to run locally next to your coding agent We built Multiplayer because we kept running into the same problem: coding agents connected to existing observability stacks inherit all the limitations those stacks were built with. Sampled traces, aggregated metrics, context that stops at service boundaries, missing request&#x2F;response content from deep within the system. The PRs they produce look plausible and fail in production (i.e. “PR slop”).<p>Multiplayer runs locally alongside Claude Code (Codex, Copilot, and Cursor coming soon) and captures full-stack, unsampled session data across your entire system. We collect everything from frontend user actions to backend traces and logs, including request&#x2F;response content and headers. It’s all the things most observability tools either sample out or don&#x27;t capture at all. We only save data when something goes wrong, so you&#x27;re not paying to store everything your system produces around the clock.<p>When an issue is identified, Multiplayer deduplicates it locally before anything reaches your coding agent. The same bug appearing across a hundred sessions becomes one issue, one prompt, one PR. Your agent works from a complete, correlated picture of what actually broke rather than a partial signal from a sampled trace.<p>We tried to make it as easy as possible to get started, so it’s just one command line to install:<p>`npm install -g @multiplayer-app&#x2F;cli &amp;&amp; multiplayer`<p>Happy to get into the architecture, the data model, or how we handle the local-first approach to data privacy.

[DevOps] Show HN: Beacon CLI for self-hosted monitoring, remote access and deployments I&#x27;ve been building a cli for my homelab&#x2F;self-hosted setup. The original motivation was getting tired of stitching together deployments, monitoring, remote SSH access and random scripts. It’s a open source CLI, all-in-one tool for monitoring, secure remote accces (tunnel and terminal), log forwarding, automated deployments. No exposed ports, or endpoint, everything is configurable how the user wants it. One use case that multiple friends are interested in is replacing HomeAssistant’s Nabu Casa subscription with tunneling through BeaconInfra. The cloud&#x2F;control-plane part (BeaconInfra) is optional. The agent itself is intended to stay local-first and continue functioning offline. Example CLI output:<p>⬡ beacon 0.5.2 ● master running uptime 29d 14h<p>DEVICE bajszi-MINI-S amd64 Ubuntu 25.04<p>PROJECTS 2 healthy 0 warning 0 down<p><pre><code> ● beaconinfra 1&#x2F;1 checks passing ● mestertkeresek 1&#x2F;1 checks passing </code></pre> TUNNELS<p><pre><code> ● homeassistant connected</code></pre>

[CLI Tool] Tsplat – Run Gaussian splatting in your terminal

[Other] Official Compound Engineering plugin for Claude Code, Codex, Cursor, and more

[IDE/Editor] Show HN: Sweetpad – VS Code extension for building iOS/Swift apps

[Other] Show HN: Iron – A Programmable Compiler that can convert 6.3M LOC/s

[CLI Tool] Show HN: VAEN – Package and import portable AI coding-agent Harnesses Hi HN, I built VAEN (an open source CLI) because I kept running into a boring problem with AI coding-agent workflows: the setup becomes useful, but then it is hard to move.<p>A good, useful agentic harness consists of more than just instructions: skills, mcp servers and more. There should be a better way to share those than just .MD files, and that is why I created VAEN.<p>What works: create a yaml, run the CLI commands as per the repo, and get a .agent file that you can share and extract. Think of how awesome it could be if anyone could use a very useful agentic harness, and share it with one CLI command.

[Other] Show HN: Open-Source AI Racing Harness Hi I&#x27;m Dan from Elodin, making an open source real-time capable flight software simulation.<p>For AI Grand Prix contestants, the wait for the Round 1 virtual qualifier simulation has been grueling.<p>If you’re competing, check out our simulation harness to tide you over, built to match the published competition constraints and message format. It runs against real Betaflight, which we learned requires at least 1000 sensor samples per second to run real-time correctly.<p>The competition warranted introducing a new feature to generate the camera sensor directly in the simulation loop. Typically people connect to Unreal or similar game engine to create a camera sensor, which works well but is very heavy. For the simple needs of this challenge, creating sample directly in the loop is very handy and easy to use. Happy to hear your feedback on this! While it&#x27;s not fancy looking currently, it uses the Rust Bevy game engine, which should allow us to improve the visual fidelity quickly.<p>We all should easily be able to shift our implementation to the published competition sim once it lands. Hope you enjoy and good luck!

[Other] Claude Code Dedicated Development Harness - Achieving High-Quality Development Through an Autonomous Plan→Work→Review Cycle

[Other] Show HN: Hodor – a 701KB native macOS prompt launcher for AI tools Hodor is a tiny macOS app that launches saved AI prompts into any text field — from the screen edge, keyboard shortcut, or keyword such as ;git.<p>I work with different AI tools every day, and had prompts scattered across Raycast snippets, Apple Notes, and Notion — notes that kept getting longer and unmanageable. Raycast snippets are useful, but cumbersome to browse and edit. I wanted one local place to save and review them, and one click to paste them into whatever AI tool I&#x27;m using. The test I set was whether I could actually stop using Raycast snippets for this. I think I fulfilled my goal. Hodor has been my daily tool for 3 months now.<p>The app is 701 KB — SwiftUI + SwiftData, no web views. Zero network requests anywhere in the code: no analytics, no telemetry, no update checks. You can verify - search the source for URLSession, it&#x27;s not in code.<p>Runs on macOS 15+, with native Liquid Glass on macOS 26+. Free and open source.<p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;woody-design&#x2F;hodor" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;woody-design&#x2F;hodor</a><p>Let me know if you have any suggestions — I&#x27;d love to hear how you solved the scattered-prompts problem.

[Other] Show HN: Filemat – an open-source web-based file manager Hello HN,<p>I would like to share Filemat, a web-based file manager that I built because I wanted something with a simple setup and file permissions that work across the filesystem (as opposed to permissions only for a folder managed by the app).<p>It&#x27;s self-hosted and open-source (currently in beta).<p>I&#x27;d be happy to hear your feedback<p>Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;bingud&#x2F;filemat" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bingud&#x2F;filemat</a>

[Other] Show HN: Game Boy pixel pipeline explorer I made a pixel pipeline explorer for the original Game Boy&#x27;s Pixel Processing Unit (PPU). If you are implementing a Game Boy emulator or just interested in it then this might provide some help :)

[Other] Incident with Pull Requests, Issues, Git Operations and API Requests

[Other] XLIDE: VBA without excel

[Other] Show HN: Gave Claude Code ADHD.. Now it thinks 3x better

[Other] Claude Code as a Daily Driver: Claude.md, Skills, Subagents, Plugins, and MCPs