{"generated_at":1789197243,"since":0,"count":173,"repos":[{"full_name":"NTBooks/jay-dee","url":"/r/NTBooks/jay-dee","github":"https://github.com/NTBooks/jay-dee","title":"jay-dee","tagline":"Your Jellyfin music library as a radio station with an AI DJ that only plays what you already own.","source":"marker","tier":"submitted","license":"MIT","stars":0,"language":"JavaScript","score":1,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"works-on-my-machine","category":["media","agent","web-app"],"built_with":["claude-code"],"slopbucket":["media","vibe-coded"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"## A DJ for the music you already have\n\nType \"rainy afternoon vaporwave with a Celtic detour\" and get ten tracks from your own Jellyfin library, in an order that makes sense, with a synthesized DJ who knows why each one is there. Call in mid-set to steer it. Thumbs down an artist and it never comes back.\n\nUnder the hood: the library is mirrored into SQLite, every artist, album and track is identified against MusicBrainz (ID3 tags are treated as hints, never truth), Claude Code subagents research every act you own, the research is embedded for semantic search, and a cheap OpenRouter model builds each set from candidates it is not allowed to invent. Playback is Webamp (Winamp 2) in the browser, or a full-art TV mode. Works on phones, deploys to Coolify.\n\nThe honest cost: the station runs on pennies with a hard daily cap, but researching a whole library burns Claude subscription usage up front. It is built to run unattended in the hours your allowance would otherwise expire, stop cleanly at a limit, and never redo anything.","readme":"Jay Dee Your Jellyfin music library as a radio station with an AI DJ. Jay Dee mirrors a Jellyfin music library into SQLite, identifies every album, artist and track against MusicBrainz (your ID3 tags are treated as hints, never truth), has Claude research every act you own, embeds the research for semantic search, and puts an AI DJ on top: you type a theme, it builds a set from your library, writes and voices the patter, and plays it in a browser station with a Winamp-style Radio mode and a full-art TV mode. Works on phones too, and deploys to Coolify. Ask for \"rainy afternoon vaporwave with a Celtic detour\" and you get ten tracks you already own, in an order that makes sense, with a DJ who knows why each one is there and what to say about it. Call in mid-set to steer it. Thumbs down an artist and it never comes back. What you need Required Notes Jellyfin yes Any recent server with an API key. The browser never talks to Jellyfin directly; audio and art are proxied. Node 22+ yes Runs the pipeline on your workstation and the station wherever you like. Claude Code yes, for research The skills in .claude/skills/ drive setup, identification, research and the DJ. See the cost section before you start. OpenRouter key for the DJ The planner runs on cheap models (gpt-5-mini / gpt-5-nano by default). Fractions of a cent per set, hard daily cap in code. A voice for the DJ Kokoro-FastAPI (free, CPU is fine) or OpenAI TTS via the same env vars. Coolify optional For running the station on ","listed_at":1789156083,"md_updated_at":1789156083},{"full_name":"NTBooks/QueryQuery","url":"/r/NTBooks/QueryQuery","github":"https://github.com/NTBooks/QueryQuery","title":"QueryQuery","tagline":"Local, heuristic triage for literary query letters. A slush-pile ticketing system that needs no AI (a local model is optional).","source":"marker","tier":"submitted","license":null,"stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"heavy","status":"works-on-my-machine","category":["productivity","web-app","data"],"built_with":[],"slopbucket":["web-app","dashboard"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"A helpdesk-style triage board for a literary agent's slush pile. Point it at a folder of query-letter `.eml` files and every letter is split into its parts, scored by a transparent heuristic engine you can edit, and laid out on a Kanban board with score-band swimlanes.\n\nBuilt for agents who are skeptical of AI, so the AI is optional: a local model can add summaries and a second opinion but never overrides the rules. Ships with 120 synthetic sample letters so you can try it without real queries. The commit history is a string of slop puns; the disclosure above is the honest version.","readme":"QueryQuery A local, helpdesk-style triage tool for the literary slush pile. Point it at a folder of query-letter .eml files and it splits each letter into its components, scores it with a transparent heuristic engine — no AI required, and lays the results out on a Kanban board with score-band swimlanes. Built for agents who are skeptical of AI: every score is explainable, the rules are yours to edit, and a local LLM (via LM Studio) is strictly optional — it only adds plain-English summaries and an advisory second opinion, and it never overrides the heuristics. Quick start Prerequisites: Node.js 22.13+ and pnpm (run corepack enable if you don't have pnpm). No native build step — the database uses Node's built-in SQLite (node:sqlite), so there's nothing to compile. git clone <this-repo> && cd queryquery cp .env.example .env # sets PERSIST_DIR (where the db + input/ live); defaults to ./persist pnpm install # installs client + server deps pnpm start # builds the client, finds a free port, starts the server, opens the browser That's the whole \"download and run.\" On first start the terminal prints the local URL and a one-time admin username + password — copy them to sign in. (Set ADMIN_PASSWORD in .env before the first run to choose your own.) Want sample data to explore? Either drag the bundled sample-data.zip letters onto the window (see Try the demo below), or generate a corpus: pnpm gen-samples # generate ~520 self-labeled sample letters (offline, no AI) pnpm scr","listed_at":1789077462,"md_updated_at":1789089375},{"full_name":"NTBooks/Fanad","url":"/r/NTBooks/Fanad","github":"https://github.com/NTBooks/Fanad","title":"Fanad","tagline":"A local-first, local-LLM RAG life-OS. You text your future self; it stays grounded in your own data.","source":"marker","tier":"submitted","license":"MIT","stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"heavy","status":"beta","category":["productivity","bot","web-app","ml"],"built_with":["claude-code"],"slopbucket":["web-app","bot","home-automation"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"You text short snippets to your future self over web chat, Telegram or Slack. A local model you run yourself (LM Studio or Ollama) classifies them into tasks, notes, lists, timers and journal entries, and every suggestion is grounded in your own rows. The model ranks and phrases; it never invents. It addresses you as PastSelf.\n\nIn daily use by its author, heavily tested, single maintainer, expect rough edges.","readme":"Fanad A local-first, local-LLM RAG life-OS. You text short snippets to your \"future self\" (web chat, Telegram, or Slack); a local LLM you configure (LM Studio or Ollama) classifies them, and every suggestion is grounded in your own data — the model only ranks and phrases real rows it's given, never invents. It addresses you as PastSelf. Status: functional and heavily tested, in daily use by its author and under active development. Single maintainer — expect fast iteration and the occasional rough edge. Prerequisites Node 24+ (uses the native built-in node:sqlite — no C++ addon; unflagged on Node 24). A local model server: LM Studio (the easy default; Developer tab → Start Server) or Ollama, with a chat model and an embedding model loaded. No model hardware? See the mock provider below, or the triple-locked cloud option under Privacy. Setup One-liner (technical users, any platform): npx github:NTBooks/Fanad This copies the app into ./fanad (no clone needed — --dir <path> to change), opens the browser setup wizard, installs dependencies, builds the web UI, and starts the server. Inside a checkout the same CLI works as npx fanad (wizard if .env is missing, then start), npx fanad setup, or npx fanad start. Windows installer (no Node needed): download FanadSetup-<version>.exe from the releases page and run it. The installer isn't code-signed yet, so SmartScreen shows an \"unknown publisher\" warning — click More info → Run anyway. It installs Fanad under %LOCALAPPDATA%\\F","listed_at":1789077457,"md_updated_at":1789089371},{"full_name":"NTBooks/clockwork-rush","url":"/r/NTBooks/clockwork-rush","github":"https://github.com/NTBooks/clockwork-rush","title":"clockwork-rush","tagline":"Gear-matching puzzle game. Originally a 2013 Windows Store (WinRT/XAML) app by the same author, now ported to the browser.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"C#","score":0,"critic_up":0,"ai_generated":"partly","human_touch":"heavy","status":"beta","category":["game"],"built_with":["claude-code"],"slopbucket":["game","ports"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"I wrote Clockwork Rush in 2013 as a Windows 8 Store app in C# and XAML. Windows 8, that Store and WinRT are all gone, so in 2026 I had an agent port it to the browser: same engine logic, same staggered gear lattice, same procedural gear geometry, same block-letter WIN and LOSE banners, in JavaScript, DOM and CSS. The original C# sits untouched next to the port.\n\nThe human wrote the game. The AI wrote the port. Both are in the repo.","readme":"Clockwork Rush A gear-matching puzzle game. Land a falling gear so its colours line up with the gear beneath it and both grind away — then rotate the whole board a quarter turn and do it again. Play it: https://clockwork-rush.ping6163.workers.dev History I wrote Clockwork Rush in 2013 as a Windows 8 Store app — WinRT and XAML, C#, with the gears drawn as procedural vector PathGeometry rather than sprites. It shipped on the Windows Store under my publisher name, Dinetopia. I'm the original author. Windows 8, the Store as it existed then, and WinRT itself have all long since moved on, and the game went with them — so in 2026 I had it ported to the browser. The original C# is still in this repo, untouched, next to the port. The port is deliberately faithful. It's the same engine logic, the same staggered gear lattice, the same collision and match rules, the same procedural gear geometry, and the same block-letter WIN / LOSE banners — just expressed in JavaScript, DOM and CSS instead of C# and XAML. Even the tutorial screenshots are the 2013 originals. It was built for Metro The look is not incidental — it's Windows 8's design language, which Microsoft called Metro until it dropped the name shortly before Windows 8 shipped. Flat colour, no gradients or bevels, content instead of chrome, and big light-weight Segoe UI. The source is full of fossils from it: every page declares the four Win8 view states — FullScreenLandscape · Filled · FullScreenPortrait · Snapped — because an app c","listed_at":1789077454,"md_updated_at":1789089364},{"full_name":"NTBooks/diet_app","url":"/r/NTBooks/diet_app","github":"https://github.com/NTBooks/diet_app","title":"diet_app","tagline":"Dead Weight — a local diet, weight and blood-pressure tracker that runs as an Electron app or a plain web page.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"works-on-my-machine","category":["productivity","data"],"built_with":["cursor"],"slopbucket":["web-app","weekend-project"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"Meal templates with calories and serving sizes, a daily log with quick-add chips for the foods you actually eat, a weight trend chart, blood pressure readings with categories, closed days, and a calendar of totals. Everything lives in a local SQLite file, with JSON export and import so the data is yours.\n\nIt is a tracker, not a doctor. The blood pressure categories are the standard bands, nothing more.","readme":"Dead Weight A comprehensive diet tracking application with meal logging, weight tracking, and blood pressure monitoring. Features 🍽️ Meal Management: Add and manage meal templates with calories and serving sizes 📝 Daily Logging: Log meals eaten and quick add items ⚖️ Weight Tracking: Daily weight logging with history ❤️ Blood Pressure: Monitor blood pressure readings with categorization 📊 Analytics: View daily calorie totals and calendar views 💾 Local Storage: All data stored locally in SQLite database Installation Prerequisites Node.js (v16 or higher) npm or yarn Development Setup Clone and install dependencies: git clone <repository-url> cd diet_app npm install Install Electron dependencies: npm install Run in development mode: npm run dev This will start both the backend server and Electron app. Production Build Build the application: npm run dist Find the built application in the dist folder Usage Starting the App Development: npm run dev (starts server + Electron) Server only: npm run server (for web browser access) Production: Run the built executable from dist folder Database The app uses SQLite for local data storage: Database file: diet_app.db Automatically created on first run No external database setup required Project Structure diet_app/ ├── main.js # Electron main process ├── index.js # Express server ├── public/ # Frontend files │ ├── index.html # Main HTML │ ├── app.js # Frontend JavaScript │ └── icon.svg # App icon ├── package.json # Dependencies and","listed_at":1789076791,"md_updated_at":1789089381},{"full_name":"NTBooks/bible-skill-public","url":"/r/NTBooks/bible-skill-public","github":"https://github.com/NTBooks/bible-skill-public","title":"bible-skill-public","tagline":"Skills vs ad-hoc tools vs training data — three ways to make an AI quote the Bible word for word, scored and appraised.","source":"marker","tier":"submitted","license":null,"stars":1,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"heavy","status":"stable","category":["docs","education","ml"],"built_with":["cursor","claude","chatgpt"],"slopbucket":["research","weekend-project"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"A small comparative study. Same task three ways: pre-built reusable scripts (a skill), tools the model writes for itself from an empty folder, and pure recall from training data. Trial one is the KJV alone; trial two quotes every verse side by side in six translations. Every quote was checked character by character against the source files.\n\nThe short version: all three ace the KJV because the KJV is everywhere, and pure memory collapses the moment the task needs six editions. The sermons, the appraisals and the scorecards are all in the repo.","readme":"AI Sermon Generation: Skills vs Ad-Hoc Tools vs Training Data A comparative study of three approaches to grounded text generation, using the Bible as a test corpus. Hypothesis Pre-built, reusable tool pipelines (skills) will outperform both AI-generated ad-hoc tools and pure training-data recall on tasks requiring exact quotation from source documents — and the advantage will grow as the task scales (more documents, more cross-referencing). Specifically: Skills will use fewer tokens because the AI never spends tokens writing code — it only calls existing scripts and composes from results. Skills will be faster on repeated tasks because indexes are built once and reused, while ad-hoc tools must bootstrap every session. Ad-hoc tools may produce better-fit tooling for novel or messy formats, since the AI adapts on the fly. Training-data recall will work for well-known texts but will fail for private documents, lesser-known sources, or tasks requiring specific editions. The gap between approaches will widen as specificity increases — more translations, stricter quoting requirements, and less common source material will expose the limits of memory-based approaches. Methodology The Task Write a short sermon about not being attached to digital things, as if they are possessions. Every Bible verse quoted must be word-for-word from the source text files. List all cited passages at the end. The Source Bible text files downloaded from openbible.com/texts.htm. Each file is tab-separated:","listed_at":1789076788,"md_updated_at":1789089379},{"full_name":"NTBooks/fanad-ha","url":"/r/NTBooks/fanad-ha","github":"https://github.com/NTBooks/fanad-ha","title":"fanad-ha","tagline":"Home Assistant App (add-on) for Fanad — local-first scratchpad","source":"marker","tier":"submitted","license":null,"stars":0,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"maintained","category":["plugin","iot","productivity"],"built_with":["claude-code"],"slopbucket":["home-automation","web-app"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"The Home Assistant add-on repository for Fanad, the local-first scratchpad for tasks, notes, lists, timers and light journaling. Add this repo's URL in the add-on store, install, and Fanad appears in your Home Assistant sidebar via ingress. Data stays in SQLite on your box; a local LLM does the parsing.\n\nThe app itself lives in [NTBooks/Fanad](https://github.com/NTBooks/Fanad). This repo holds the add-on packaging, release notes and the fanad.org landing page.","readme":"Fanad — Home Assistant add-on repository This repository lets you install Fanad as a Home Assistant App (add-on): a local-first personal scratchpad for tasks, notes, lists, timers and light journaling. Your data stays in SQLite on your box; a local LLM (LM Studio / Ollama) does the parsing. No nagging, no cloud by default, no telemetry. Install In Home Assistant go to Settings → Add-ons → Add-on store (on newer installs: Settings → Apps). Open the ⋮ menu (top right) → Repositories. Add this URL: https://github.com/NTBooks/fanad-ha Find Fanad in the store, click it, and Install. Start it, open the Web UI, and connect a model in Settings (see the add-on's Documentation tab). Prefer plain Docker? The same image runs anywhere: docker run -p 8787:8787 -v fanad:/data ghcr.io/ntbooks/fanad-app:latest","listed_at":1789076781,"md_updated_at":1789089368},{"full_name":"NTBooks/a2ha","url":"/r/NTBooks/a2ha","github":"https://github.com/NTBooks/a2ha","title":"a2ha","tagline":"Assistant to the Home Assistant - a Pinata agent that runs your house from chat and hands guests a locked-down button pad through an expiring link","source":"marker","tier":"submitted","license":"MIT","stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"alpha","category":["agent","iot","automation"],"built_with":["claude-code"],"slopbucket":["agent","home-automation"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"Someone is staying at your place for a week. They need the porch light, the guest room lamp and the coffee machine, and nothing else. A2HA makes them a pad of big buttons behind an expiring link, tests that each button actually moves something, and kills the link when they leave. From chat, in about four seconds.\n\nEvery commit is co-authored by Claude Opus 5.","readme":"A2HA — Assistant to the Home Assistant Not the Assistant Home Assistant. That's a different and frankly less prestigious role, and the distinction matters to this agent more than it does to anyone else. A2HA is a Pinata agent that does two things for a Home Assistant house: it hands guests a strictly limited set of buttons for exactly as long as they're staying, and it does the fiddly configuration work you keep putting off. See the overview page → — a one-page explanation you can send to someone who hasn't read any of this. The problem it actually solves Someone is staying at your place for a week. They need the porch light, the guest room lamp, and the coffee machine. Your options today are all bad. Make them a Home Assistant account and they can see every camera, every door lock and every sensor in the house — and you have to remember to delete it afterwards. Hand them your phone. Or leave a note and hope. A2HA gives you a third option: a link. you: make a pad for the house-sitter with the porch light, the guest room lamp and the coffee machine a2ha: Done — three buttons. I tested all three, they actually moved. Want a link? I'd go 7 days. you: yeah, they arrive Friday a2ha: https://your-agent.agents.pinata.cloud/pad/r/a2h1_… Porch light, Guest room lamp, Coffee machine. Dies next Sunday. Copy it now — I can't show it again. They open it on their phone. Three big buttons. No app, no account, no login. When the link expires it stops working, and you can kill it earlier from","listed_at":1789076770,"md_updated_at":1789089361},{"full_name":"NTBooks/context-game","url":"/r/NTBooks/context-game","github":"https://github.com/NTBooks/context-game","title":"context-game","tagline":"Vibe Blaster! Learn to manage an LLM's context window by blasting vibes in a turn-based pixel-art lane shooter.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"beta","category":["game","education"],"built_with":["claude-code"],"slopbucket":["game","vibe-coded"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"Every shot you charge costs context. Sit in the 50–70% band and prompts hit hard; over-explain and you overkill, waste tokens and creep toward 85% where the hallucination effects kick in. Cross 95% and the window compacts: enemies jump forward and permanent sludge fills the bottom of the bar. Split the work to a ghost tank, heal, rewind, or end your turn and let the wave advance.\n\nAll sprites, shaders and sound are generated at boot. The `scripts/` folder holds the image-generation helpers used to draft the art.","readme":"Vibe Blaster! Learn how to manage your LLM context by blasting vibes. Play it: https://contextgame.dsoul.org A turn-based pixel-art lane shooter built with Phaser 3. You are a tank defending the codebase. Enemies (bugs, vibes, and the occasional boss) step down three lanes toward the fortress. Every prompt you fire is charged by hand, and every charge spends context. The mechanic Context window replaces heat. The system prompt permanently occupies the bottom 20%. Charge a shot and stop inside the drifting perfect zone for a damage bonus. The zone moves and changes width every shot. Stay in the 50–70% band. That is the optimal range. Overkill (killing something with 2x its remaining HP) wastes context, because over-explaining a simple task is expensive. 85% and hallucination effects start. 95% triggers compaction: enemies jump forward two steps and permanent sludge fills the bottom of the bar. Abilities: split the work to a ghost tank, heal, rewind, charge, cost tricks. Buy them in the shop between waves. End turn and the wave advances. Touch controls work on phones. Run locally It is a static site. Serve the folder with anything: npx serve . Phaser is loaded from a CDN. All sprites, shaders and sound are generated programmatically at boot; the assets/ folder holds the skill icons and logos/ the branding. The scripts/ folder contains the AI image-generation helpers used while drafting the art (npm run gen, gen-bg, rm-bg). They are not needed to play. Slop disclosure This repo ","listed_at":1789076766,"md_updated_at":1789089355},{"full_name":"NTBooks/gpt5-test","url":"/r/NTBooks/gpt5-test","github":"https://github.com/NTBooks/gpt5-test","title":"gpt5-test","tagline":"VaporVU — a vaporwave 3D music visualizer, vibe-coded in an afternoon to see what GPT-5 could do without anyone reading the code.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"entirely","human_touch":"none","status":"prototype","category":["media","art","toy"],"built_with":["cursor","chatgpt"],"slopbucket":["vibe-coded","weekend-project","music"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"The commit log says it best: \"Vibe coded, didn't even look at it.\"\n\nDrop MP3s in `visualizer/public`, run `npm run dev`, and a neon skyline of frequency bars, a mirror floor, bloom, afterimage trails and popper fireworks react to the music. A second folder holds an unrelated mock of a student credential profile page from the same test session.","readme":"gpt5-test Two things vibe-coded in August 2025 to see what GPT-5 could do when nobody reads the output. Per the commit log: \"Vibe coded, didn't even look at it.\" visualizer/ — VaporVU A vaporwave 3D music visualizer built on three.js, Vite and the Web Audio API. Frequency bars rise out of a mirrored floor under a neon sky, with bloom, afterimage trails, wind-blown smoke, embers and popper fireworks that fire on the beat. Orbit, follow and centre-up cameras, performance presets, and a screenshot button. cd visualizer npm install npm run dev # http://localhost:5173 Drop .mp3, .wav, .flac or .m4a files into visualizer/public/ and they appear in the playlist. Three sample tracks are included. mock/ — Student Credential Profile An unrelated static mock of a student credential profile page (Tailwind, no build step) from the same test session. Open mock/index.html in a browser. Slop disclosure This repo is listed on SlopScore. See slopscore.md for the AI-generation disclosure.","listed_at":1789076763,"md_updated_at":1789089352},{"full_name":"NTBooks/Cherry","url":"/r/NTBooks/Cherry","github":"https://github.com/NTBooks/Cherry","title":"Cherry","tagline":"雪月花 — a cel-shaded, looping 3D screensaver of a samurai's grave as four seasons turn and the cherry trees bloom.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"works-on-my-machine","category":["art","toy","media"],"built_with":["cursor"],"slopbucket":["vibe-coded","weekend-project","screensaver"],"has_images":true,"vulns":{"deps":0,"vulnerable":0},"pitch":"A katana stands in a stone-fenced cherry grove and the camera keeps a slow vigil while summer, autumn, winter and spring roll past. Everything is procedural: no models, no textures, no downloads. Three.js is vendored so it runs offline.\n\nPress `4` to skip straight to the petal storm.","readme":"Cherry 雪月花 — Sakura Snow Globe. A cel-shaded, looping 3D screensaver that runs as a web page: a katana stands at a samurai's grave in a stone-fenced cherry grove, and the camera keeps a slow vigil while the seasons turn. Summer, autumn leaf-fall, bare winter snow, spring cherry blossoms, and around again. Live: https://cherryfable.pages.dev Everything is procedural. No models, no textures, no downloads. Three.js r184 is vendored, so the whole thing works offline. Run it locally The screensaver lives in snowglobe/. It uses ES modules, so it needs an HTTP server rather than a file:// open. Windows: double-click snowglobe/Start Screensaver.cmd. It starts a tiny PowerShell server on port 8423 and opens your browser. Anything else: cd snowglobe python -m http.server 8423 # or: npx serve -p 8423 then open http://localhost:8423/. For a chromeless window try msedge --app=http://localhost:8423/. Controls Input Effect nothing slow orbit, four day/night cycles per season, loops forever drag orbit the camera scroll dolly in and out click a gust of wind stirs petals, leaves, rain and snow double-click fullscreen 1 2 3 4 jump to summer / autumn / winter / spring + / − time speed R render resolution C real-time clock mode Space pause the weather and particles for screenshots The full feature list, tuning knobs and file layout are in snowglobe/README.md. Slop disclosure This repo is listed on SlopScore. See slopscore.md for the AI-generation disclosure.","listed_at":1789076759,"md_updated_at":1789089348},{"full_name":"NTBooks/slopscore","url":"/r/NTBooks/slopscore","github":"https://github.com/NTBooks/slopscore","title":"slopscore","tagline":"Give me your slop! Peer review for code nobody wrote.","source":"marker","tier":"submitted","license":null,"stars":0,"language":"TypeScript","score":1,"critic_up":0,"ai_generated":"entirely","human_touch":"light","status":"alpha","category":["web-app","devtools"],"built_with":["claude-code"],"slopbucket":["web-app","devtools","vibe-coded"],"has_images":true,"vulns":{"deps":172,"vulnerable":0},"pitch":"## The site lists itself\n\nSlopScore was built in one sitting by an agent, from a plan, which is exactly the kind of thing it exists to grade. Every page is also `.json` and `.md`. The queue, the moderation log, and the free-tier headroom are public.\n\nNothing here is a secret. That's the point.","readme":"SlopScore Give me your slop! Peer review for code nobody wrote. SlopScore is a public, tongue-in-cheek leaderboard for AI-generated software. A repo owner opts in by committing a slopscore.md file to a public GitHub repo. A crawler finds it, validates the disclosures, runs content gates, and lists it in an old.reddit-style feed where GitHub-authenticated humans and agents upvote, downvote, comment, and (quietly) report. Live: https://slopscore.org · staging: https://test.slopscore.org List your repo Commit this to slopscore.md at the root of your default branch: --- slopscore: 2 spec: https://slopscore.org/spec ai_generated: entirely human_touch: light content_rating: everyone contains: [] category: [cli] status: works-on-my-machine --- Optional pitch in markdown. If empty, your README is the pitch. That's the whole file. Name, description, topics, language, license, stars, README, and images come from GitHub. Impatient, or the crawler hasn't found you? curl https://slopscore.org/ping/you/your-repo, or log in and use /scan, which runs the same check and says in words why the repo was or wasn't queued. Full contract: /spec · docs/SPEC.md. Found repos are listed and votable straight away. Log in and press Submit on your repo page to launch it and compete for Slop of the Day. Principles Just add the file. No submit form, no webhook, no app to install. Store nothing we don't own. Only our database. GitHub owns identity, code, images, and the marker file. Fully transparent. Every ","listed_at":1789072854,"md_updated_at":1789088953},{"full_name":"Wentong-Liu/SayIt","url":"/r/Wentong-Liu/SayIt","github":"https://github.com/Wentong-Liu/SayIt","title":"SayIt","tagline":"FREE version of Typeless/Wisper/Monologue/闪电说 - no word limit, no subscription","source":"trawl","tier":"found","license":"MIT","stars":49,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: FREE version of Typeless/Wisper/Monologue/闪电说 - no word limit, no subscription; its own README says \"License MIT (LICENSE) Built with Claude Code (\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"SayIt Free, native macOS voice dictation with AI polish. Hold a key, speak, and clean, polished text lands at your cursor — in any app. ⬇️ Download the latest SayIt.dmg · All releases English · 简体中文 SayIt turns your voice into polished text anywhere on your Mac. It's built for people who already talk to AI all day — vibe coders, writers, anyone who would rather speak than type. Hold the key, say what you mean, and the cleaned-up text appears right where your cursor is. Why SayIt ⚡ Instant, on-device speech Your speech is transcribed locally on your Mac, accelerated by the Apple Neural Engine — the same NPU that powers on-device AI. It's quick, and it stays quick no matter how long you spoke: measured on an Apple M4 Pro with the recommended large-v3-turbo model, transcription runs in about 0.4 seconds — an ~9-second clip became text in ~0.4 s, roughly 20× faster than real time, and a short phrase finishes just as fast. Exact speed depends on your Mac, but the work always runs on-device, fully offline, and your audio never leaves your machine. On macOS 26 and later, SayIt defaults to Apple's built-in speech engine — the new on-device SpeechAnalyzer model, with nothing to download. It's faster still: transcription is effectively instant, so quick that the \"Transcribing…\" indicator barely has time to appear before your words land at the cursor. Older macOS — or your own preference — falls back to the bundled WhisperKit model above. ♻️ Uses your Codex subscription to polish If you","listed_at":1789172413,"md_updated_at":1789172413},{"full_name":"Thairu-dev/StealthIt","url":"/r/Thairu-dev/StealthIt","github":"https://github.com/Thairu-dev/StealthIt","title":"StealthIt","tagline":"An invisible AI desktop assistant that evades screen capture and taskbar detection. Features instant visual analysis, live Whisper voice transcription, and robust support for premium LLMs (Anthropic, OpenAI, OpenRouter, Gemini, and local Ollama).","source":"trawl","tier":"found","license":"MIT","stars":14,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: An invisible AI desktop assistant that evades screen capture and taskbar detection. Features instant visual an; its own README says \"- /claude-opus-5 : A complete rebuild and enhancement of the app, built by Claude Opus 5 (featuring native desktop architecture, custom Open\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"StealthIt — Invisible AI Meeting & Desktop Assistant (Free Cluely Alternative) StealthIt is a free, open-source Cluely alternative and invisible AI desktop assistant designed for live meeting notes, technical interviews & coding, instant screen analysis, and voice transcription. Floating directly over your workspace, StealthIt gives you instant access to state-of-the-art AI models (Claude, OpenAI, Gemini, Ollama)—while remaining completely hidden from the taskbar and 100% invisible to anyone viewing your screen share or recording (Zoom, Microsoft Teams, Google Meet, Discord, OBS). Whether you need a silent AI meeting assistant to summarize audio in real time, a live coding copilot that reads problem statements via OCR, or an invisible presentation prompter, StealthIt works seamlessly across all your apps without ever being captured on video calls. Built with Python and PySide6/Qt, this repository also hosts the AI Models Competition, tracking how different LLM architectures build and optimize invisible desktop software. 💼 Core High-Stakes Use Cases StealthIt functions as a versatile AI Desktop Assistant across multiple professional scenarios: Use Case Description Primary Hotkeys 🎙️ Live Meeting Assistant & AI Summarizer Silently transcribes Zoom, Teams, Google Meet, or in-person audio locally via Whisper. Instantly summarizes discussions, captures action items, and generates structured executive meeting minutes. Ctrl+R 💻 Technical Interviews & Live Coding F","listed_at":1789172125,"md_updated_at":1789172125},{"full_name":"zahhar/ghcp-dashboard","url":"/r/zahhar/ghcp-dashboard","github":"https://github.com/zahhar/ghcp-dashboard","title":"ghcp-dashboard","tagline":"Dashboard with Github Copilot Metrics","source":"trawl","tier":"found","license":"MIT","stars":14,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["productivity"],"built_with":["copilot"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Dashboard with Github Copilot Metrics; its own README says \"jpg) ⚠️ Disclaimer: this is a fully vibe-coded project that did not go through comprehensive code review or testing\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Yet Another (but cool) Github Copilot Metrics Dashboard Lightweight GitHub Copilot usage dashboard for teams. It reads raw metrics exposed via new Github REST API endpoints for Copilot usage metrics, aggregates them per user, and serves a browser-based leaderboard-style dashboard with filters and trend hints. Metrics captured Detailed documentation on every metrics The dashboard tracks activity across two Copilot interfaces: IDE plugin — chat asks, code completions, agent/coding-agent runs; favorite IDE with version, language, model CLI — GitHub Copilot CLI prompt counts, active days via CLI, and last known CLI version (shown in user detail popup) ⚠️ Disclaimer: this is a fully vibe-coded project that did not go through comprehensive code review or testing. Results may be inaccurate, and bugs are possible. Why this project exists Github Copilot built-in dashboards are still in Public Preview, they are yet rudimentary and require special access rights difficult to obtain in large enterprises. Existing external dashboards (by Github, by Microsoft) were not promptly updated for new REST API compartibility and likely stop working on April 2, 2026 when Github sunsets its legacy Github Metrics API. Target audience Dashbord aims at AI/Agile Coaches, Teams leaders, Engineering managers, Project- and Delivery managers, Procurement associates and helps themto quickly answer questions like: Who is actively using Copilot and who is not? Which models/IDEs/languages are most used? How usag","listed_at":1789172121,"md_updated_at":1789172121},{"full_name":"serrebidev/BlindPilot","url":"/r/serrebidev/BlindPilot","github":"https://github.com/serrebidev/BlindPilot","title":"BlindPilot","tagline":"Vibe coded, accessible desktop frontend for Agents like codex, Claude, etc and chatting with AI's through API's.","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Vibe coded, accessible desktop frontend for Agents like codex, Claude, etc and chatting with AI's through API'; its own README says \"Vibe coded, accessible desktop frontend for Agents like codex, Claude, etc and chatting with AI's through API's\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"BlindPilot A screen-reader-first desktop front end for AI coding CLIs. It runs Claude Code, Codex, FreeBuff, opencode, and Hermes in native wxPython windows, so NVDA, JAWS, and VoiceOver read controls instead of a terminal. It runs on Windows, macOS, and Linux. Linux is the least tested of the three. Questions, bugs, and release news go to the SerrebiProjects Telegram group. Bug reports and feature requests also go in Issues. BlindPilot started as a fork of Claude Code Reader by doubletaponair and keeps its accessibility design. See CREDITS.md. What it does Runs five coding agents, picked from Model, Backend. The choice is remembered. Runs every backend in Bypass permissions mode by default, so a task does not stop to ask for approval. Change this under Model, Permission Mode. Splits every answer into rows you can arrow through, one per heading, paragraph, list item, quote, code block, thought, tool call, and tool result. Reads answers aloud as they stream, or stays silent until the whole answer is in. Speaks every step of a run, or only your message, the answer, and status changes. See Narration below. Reopens past conversations from any backend and continues them. Compacts a long one in place. Runs several sessions at once, one tab each, with its own folder, model, and permission mode. Answers the multiple-choice questions a backend stops to ask, in one dialog with radio buttons, checkboxes, and an Other box. Lets you steer a running task with a new message, or stop it and ","listed_at":1789172118,"md_updated_at":1789172118},{"full_name":"mlemos/parker","url":"/r/mlemos/parker","github":"https://github.com/mlemos/parker","title":"parker","tagline":"The scratchpad for vibe coding — a fast, minimal, keyboard-first Mac editor for notes, code & markdown.","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["devtools"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":794,"vulnerable":8},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: The scratchpad for vibe coding — a fast, minimal, keyboard-first Mac editor for notes, code & markdown.; its own README says \"It isn't a commercial product — it's a passion project, vibe-coded for my own use and shared with the world\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"The scratchpad for vibe coding. A fast, minimal Mac editor for notes, code & markdown — the serious scratchpad for working with AI, without the weight of Notes or VS Code. NoteParker is beta, and personal. It isn't a commercial product — it's a passion project, vibe-coded for my own use and shared with the world. No guarantees, no support, use at your own risk. 🙂 What is Parker? Parker is a resident menu-bar app for macOS — summon it from anywhere with a global shortcut, jot or draft, dismiss it, and never lose a thought. It's built for the way people work with AI today: long prompts, task lists, quick snippets and throwaway notes that deserve something faster than Notes and lighter than VS Code. It operates on a plain folder of files you own. No accounts, no database, no servers of ours — ever. ✨ Features Instant & fast — opens in a blink, zero lag while typing. No splash, no spinners, no \"syncing…\". Keyboard-first — summon with a global shortcut; almost everything is reachable without the mouse. Lightweight — a tiny native shell on Tauri, around 10 MB. All editor, no bloat. Never lose data — aggressive auto-save, always on, plus optional git backup as a safety net. Split & preview — split panes and live Markdown preview, side by side. Genuinely themeable — eight themes: Vercel Night & Day, GitHub Light & Dark, plus Playa, Playa at Night, Matrix and Blueprint. Your data, your rules — plain .md, .txt and code on disk, in a folder you choose. 🔒 Your data,","listed_at":1789172114,"md_updated_at":1789172114},{"full_name":"chynggi/gigatoken-llama.cpp","url":"/r/chynggi/gigatoken-llama.cpp","github":"https://github.com/chynggi/gigatoken-llama.cpp","title":"gigatoken-llama.cpp","tagline":"LLM inference in C/C++ with Gigatoken / This Repo is Very Vibe-Coded-Experimental","source":"trawl","tier":"found","license":"MIT","stars":10,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: LLM inference in C/C++ with Gigatoken / This Repo is Very Vibe-Coded-Experimental; its own README says \"LLM inference in C/C++ with Gigatoken / This Repo is Very Vibe-Coded-Experimental\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"llama.cpp GigaToken Integration This fork integrates GigaToken, a SIMD-accelerated BPE tokenizer achieving ~1000x faster throughput than HuggingFace tokenizers. GigaToken replaces the default tokenization pipeline with a Rust-based implementation that reads data directly and maximizes parallelism, enabling tokenization at GB/s speeds on modern CPUs. Key features: Drop-in replacement compatible with existing HuggingFace tokenizer workflows SIMD-optimized pretokenization (AVX512/AVX2/NEON) Concurrent cache hierarchies for long-tailed pretoken distributions Support for all commonly used BPE tokenizers (GPT-2, Llama 3, Qwen, DeepSeek, etc.) LLM inference in C/C++ ggml / ops / maintainer PRs / dev stats / lib llama API / llama-server REST API Quick start A few options to get llama.cpp installed on your machine: Visit https://llama.app and follow the instructions Run with Docker - see our Docker documentation Download pre-built binaries from the releases page Build from source by cloning this repository - check out our build guide Once installed: # Download and run a model directly from Hugging Face llama cli -hf ggml-org/Qwen3.5-0.8B-GGUF # Launch OpenAI-compatible API server llama serve -hf ggml-org/Qwen3.5-0.8B-GGUF VLM session with llama cli Built-in web UI against llama serve DescriptionThe main goal of llama.cpp is to enable LLM (and VLM) inference with minimal setup and state-of-the-art performance on a wide range of hardware - locally and in the cloud. Plain C/C++ implement","listed_at":1789171822,"md_updated_at":1789171822},{"full_name":"MarkMakies/anthropic-explorer","url":"/r/MarkMakies/anthropic-explorer","github":"https://github.com/MarkMakies/anthropic-explorer","title":"anthropic-explorer","tagline":"A PyQt6 desktop GUI for testing and exploring Anthropic's Claude API. Built with Claude Code.","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A PyQt6 desktop GUI for testing and exploring Anthropic's Claude API. Built with Claude Code.; its own README says \"Built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Anthropic API Explorer A desktop GUI for testing and exploring Anthropic's Claude API. Built with PyQt6, designed for full visibility of every message type, content block, parameter, and cost. Built entirely using Claude Code as a personal project to explore both the Claude API and the experience of building software collaboratively with AI. Note: This is an independent personal project. I have no affiliation with Anthropic, I'm just a private customer who wanted a better way to explore their API. Features All Claude models — auto-refreshed model list with manual fallback Streaming and classic modes — watch responses arrive in real time or wait for the full result Extended thinking — enable thinking with configurable budget; view reasoning in a dedicated tab Web search — built-in web search tool support with citation rendering Full parameter control — temperature, top-p, top-k, max tokens, timeout 6 output tabs: Conversation (with optional markdown rendering) Thinking (extended reasoning log) Request JSON (syntax-highlighted) Response JSON (syntax-highlighted) Usage & Cost (persistent log across sessions) Raw Blocks (every content block type verbatim) Conversation management — multi-turn and single-turn modes, import/export to JSON Cost tracking — per-request and cumulative session costs, saved to usage_log.txt across sessions Markdown toggle — render **bold**, *italic*, `code`, headers, lists in the conversation view Screenshot Prerequisites You'll need an Anthropic API ","listed_at":1789171819,"md_updated_at":1789171819},{"full_name":"chatelp/geek-casino-cardputeradv","url":"/r/chatelp/geek-casino-cardputeradv","github":"https://github.com/chatelp/geek-casino-cardputeradv","title":"geek-casino-cardputeradv","tagline":"Silicon Casino — five-game offline casino for M5Stack Cardputer ADV (ESP32-S3). Slots, blackjack, video poker, roulette. Virtual chips only — a toy, built with Claude Code.","source":"trawl","tier":"found","license":"MIT","stars":11,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Silicon Casino — five-game offline casino for M5Stack Cardputer ADV (ESP32-S3). Slots, blackjack, video poker,; its own README says \"Virtual chips only — a toy, built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"English · Français Silicon Casino A five-game casino in your pocket — virtual chips only. M5Stack Cardputer ADV · ESP32-S3 · 240 × 135 display · fully offline. Shake to spin. Silicon Casino is an offline casino toy for the M5Stack Cardputer ADV: three-reel slots, a 5×3 video slot, blackjack, Jacks-or-Better video poker and a European roulette — sharing one balance, one leaderboard, and one rule: virtual chips only. No real money, no purchases, no account, no network. You can never be ruined for good: the house refills you. The randomness comes from the ESP32's hardware generator. Its visual identity is the machine itself: the slot cabinet is a printed circuit board (mounting holes, traces, vias), the blackjack table is routed like one, the roulette wheel sits on a rotary encoder — and the reels spin geek glyphs. A setting brings the classic symbols back: the geek skin is a gift, not a wall. Getting around Everything is played one-handed, and the lobby tells you where you are. Key Action space / enter confirm — spin, deal, draw, launch the ball shake the device pulls the lever: spins the reels, launches the ball ← → change bet — or move the cursor in a hand ↑ ↓ navigate the lobby, pick a roulette number 1…9 0 house credit — +10 to +90, 0 adds +100 · anywhere, anytime h help — general in the lobby, the game's rules in a game s settings — global in the lobby, per-game in a game l leaderboard a about esc back — works everywhere, any time, even mid-hand esc (lobby) quit On the dev","listed_at":1789171816,"md_updated_at":1789171816},{"full_name":"MikeWilson/esp32-gameos","url":"/r/MikeWilson/esp32-gameos","github":"https://github.com/MikeWilson/esp32-gameos","title":"esp32-gameos","tagline":"A handheld game OS + six procedural games for the Waveshare ESP32-S3 Touch AMOLED 1.8 — built with Claude Code","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A handheld game OS + six procedural games for the Waveshare ESP32-S3 Touch AMOLED 1.8 — built with Claude Code; its own README says \"8 — built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"gameos A handheld game OS — and a growing set of games — for the Waveshare ESP32-S3-Touch-AMOLED-1.8 (368x448 AMOLED, touch, IMU, speaker, 8 MB PSRAM). Plain C on ESP-IDF; no game engine, no LVGL in the frame path, no sprite/PCM asset files. Everything you see and hear is procedural. Launcher Gunship Golf Golf: aiming (Frames rendered by the host simulator — pixel-identical to the device.) The games GUNSHIP — AC-130 thermal gunner over a dead town. Tilt to aim, time-of-flight shells against an orbiting camera, 10 waves, upgrades. Indexed-palette thermal rendering with scanlines and sensor noise. GOLF — a full port of the standalone Infinite Golf: procedural SDF-generated courses, three clubs, wind, swing-by-actually-swinging-the- device (IMU stroke integration), 18-hole rounds, pass-and-play for 2-4, leaderboard. Full-res RGB565 with real Montserrat type. SLOTS — a palette-trickery slot machine with a physical pull gesture. DIAG / AIM TEST — built-in hardware diagnostics and the tilt-aim latency test harness. Architecture games → contract (gos.h) → shell → core → HAL (dependencies point down) A game includes only gos.h and registers one struct: id, capability flags, a state size, and five function pointers (init/update/render/ suspend/teardown). Adding a game = one registry line + a component. All games compile into the single firmware image and execute in place from flash. \"Launching\" a game is a calloc of its declared state (internal SRAM, PSRAM fallback for the multi-megab","listed_at":1789171812,"md_updated_at":1789171812},{"full_name":"4uffin/auto-news-aggregator","url":"/r/4uffin/auto-news-aggregator","github":"https://github.com/4uffin/auto-news-aggregator","title":"auto-news-aggregator","tagline":"An automated tech news aggregation and summarization tool that delivers concise TL;DR summaries from major tech publications every 12 hours.","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["automation"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tech news digest that summarizes major tech publications every 12 hours, whose README says \"THE REST OF THIS PROJECT WAS CREATED USING GOOGLE GEMINI 2.5 FLASH.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"WarningTHIS README WAS GENERATED BY CLAUDE SONNET 4 AND HAS BEEN DOUBLE CHECKED BY 4UFFIN. THE REST OF THIS PROJECT WAS CREATED USING GOOGLE GEMINI 2.5 FLASH. EXPECT BUGS AND USE WITH CAUTION! Tech News Digest 📰🤖 An intelligent, automated tech news aggregation and summarization tool that delivers concise TL;DR summaries from major tech publications every 12 hours. 🚀 Quick Start # Clone and setup git clone <your-repository-url> cd tech-news-digest pip install -r requirements.txt # Set API key export OPENROUTER_API_KEY=\"your-openrouter-api-key\" # Generate your first digest python generate_digest.py ✨ Features Feature Description 🌐 Multi-Source Aggregation Pulls from 5+ major tech publications 🧠 AI-Powered Summaries Uses Google Gemini 2.5 Flash for intelligent summarization 🔄 Robust Fallbacks RSS feeds → Web scraping → Error handling ⏰ Automated Scheduling Runs every 12 hours via GitHub Actions 📁 Organized Storage Timestamped digests with proper file structure 🔧 Easy Configuration Simple customization of sources and parameters 📊 News Sources Publication RSS Feed Backup Scraping 📱 The Verge ✅ ✅ 💼 TechCrunch ✅ ✅ 🔧 CNET ✅ ✅ ⚗️ Ars Technica ✅ ❌ 📟 Engadget ✅ ❌ 🛠️ Installation & Setup Prerequisites Python 3.12+ - Download here OpenRouter API Key - Get yours here Git (for cloning and automation) Local Setup Clone the repository: git clone <your-repository-url> cd tech-news-digest Create a virtual environment (recommended): python -m venv venv source venv/b","listed_at":1789166932,"md_updated_at":1789166932},{"full_name":"rusiaaman/syntax-checker","url":"/r/rusiaaman/syntax-checker","github":"https://github.com/rusiaaman/syntax-checker","title":"syntax-checker","tagline":"Multiple languages syntax checker","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1000,"vulnerable":69},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Python library that checks syntax errors across many languages using Tree-sitter, whose README says \"this repo is 100% AI generated and auto tested.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Syntax-checker: A Multi-Language Syntax Checker A Python library that uses Tree-sitter to check syntax errors across multiple programming languages based on Rust. ⚠️ Warning: this repo is 100% AI generated and auto tested. However, some human review and testing has been done. Installation pip install syntax_checker Features Support for multiple languages including: Fast parsing using tree-sitter Precise error reporting with line and column numbers No need for language-specific toolchains Easy to extend with additional language support Usage The library provides a simple interface to check syntax errors in source code: import syntax_checker # Check Python syntax output = syntax_checker.check_syntax(\"py\", \"\"\" def invalid_function[T](x): # Invalid type parameter syntax print(\\\"unclosed string') # Unclosed string \"\"\") # Get error positions as (line, column) tuples print(output.errors) # [(2, 19), (3, 11)] # Get error descriptions print(output.description) # Line 2, Column 19: Syntax error: unexpected ERROR in function_definition # Line 3, Column 11: Missing string_content in string # Check a file without errors output = syntax_checker.check_syntax(\"py\", \"\"\" def valid_function(x: int) -> int: return x * 2 \"\"\") print(output.errors) # [] print(output.description) # \"\" Supported Languages and File Extensions Language Extension(s) Bash sh C c, h C++ cpp, hpp C# cs CSS css Elisp el Elixir ex, exs Elm elm Go go HTML html, htm Java java JavaScript js JSX jsx JSON json Lua lua PHP php ","listed_at":1789166930,"md_updated_at":1789166930},{"full_name":"slungus57/Kestrel","url":"/r/slungus57/Kestrel","github":"https://github.com/slungus57/Kestrel","title":"Kestrel","tagline":"Kestrel is a lightweight Roblox anti-AFK tool designed to keep your session active by simulating normal player activity. Simple, efficient, and easy to use, Kestrel helps prevent automatic AFK kicks while staying unobtrusive. This project is fully vibe coded.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Kestrel, a terminal Windows utility that sends timed key presses to keep a Roblox window from going AFK; the owner says \"This project is fully vibe coded.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Kestrel Keep your Roblox window awake. Kestrel is a terminal-only Windows utility that periodically performs configurable OS-level keyboard actions. It uses pynput for normal keyboard input and does not inject into Roblox, inspect Roblox memory, modify Roblox files, hook game internals, or manipulate network traffic. Requirements Windows 10/11 Python 3.11 or newer A normal terminal such as Windows Terminal, PowerShell, or Command Prompt Installation From the Kestrel project directory: py -3.11 -m venv .venv .venv\\Scripts\\activate python -m pip install --upgrade pip pip install -r requirements.txt Run python main.py Optional configuration path: python main.py --config path\\to\\config.json Show the configuration guide without starting the dashboard: python main.py --config-help Configuration Kestrel automatically creates config.json with safe defaults if it is missing. Example: { \"enabled\": false, \"interval\": 600, \"randomization\": 10, \"actions\": [ { \"key\": \"w\", \"duration\": 1.0 }, { \"key\": \"s\", \"duration\": 1.0 }, { \"key\": \"space\", \"duration\": 0.2 } ], \"target_window\": \"Roblox\" } enabled Boolean. When true, Kestrel starts active after initialization. When false, it starts stopped. interval Number of seconds between action cycles. 600 = every 10 minutes 300 = every 5 minutes 60 = every minute 10 = every 10 seconds randomization Non-negative number of seconds. Each interval is randomly adjusted by approximately ±randomization. 0 disables randomization. actions An ordered list. Every","listed_at":1789166927,"md_updated_at":1789166927},{"full_name":"jsglazer/annotation-manager","url":"/r/jsglazer/annotation-manager","github":"https://github.com/jsglazer/annotation-manager","title":"annotation-manager","tagline":"Obsidian plugin to create and query annotations","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":1,"critic_up":1,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an Obsidian plugin for inline annotations with a sidebar and Dataview queries, whose README carries a \"Made with Claude\" badge. It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Annotation Manager Annotate text inline in Obsidian, view all annotations for the current note in a sidebar, style them with custom colors, attach freeform tagged comments, and query annotations with Dataview. View all/some/none of the annotation markup Documentation (Wiki) Full documentation is in the GitHub Wiki: Home — overview and quick start Annotations — syntax, styling, sidebar, commands Comments — syntax, tag inheritance, sidebar, commands Configuration — all settings, including the per-theme color grids DataviewJS Examples — querying annotations A link to the wiki is also on the plugin's Settings → Annotation Manager → General tab. What it does Inline annotations — wrap any text in {={parent/child}your note=} to tag it with an identifier Live Preview rendering — delimiters and identifier are hidden; only the annotated text is shown, optionally styled Annotations sidebar — titled \"Annotations\" at the top; lists the annotations in the note you currently have open, grouped by identifier, with one-click navigation; switches automatically as you move between notes Comments — attach freeform notes with {@your comment@}, independent of annotations; the Add comment command prompts for a tag from the same list as Apply identifier to selection (pick \"No tag\" for an untagged comment), or tag one explicitly by typing {@{parent/child}your comment@}. Run it with the cursor immediately after an annotation (no space) and it skips the prompt entirely, inheriting that annotation's tag","listed_at":1789166925,"md_updated_at":1789166925},{"full_name":"dynamicheart/llm-eval-viewer","url":"/r/dynamicheart/llm-eval-viewer","github":"https://github.com/dynamicheart/llm-eval-viewer","title":"llm-eval-viewer","tagline":"A lightweight web viewer for LLM evaluation results — visualize, browse and analyze evalscope outputs instantly in your browser.","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"JavaScript","score":2,"critic_up":3,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a lightweight browser viewer for evalscope LLM evaluation results whose README says \"This project was built entirely with AI assistance\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"LLM Eval Viewer 中文文档 This project was built entirely with AI assistance (Claude Opus 4.6). LLM Eval Viewer is a lightweight web tool for visualizing LLM evaluation results. Currently supports result formats generated by evalscope. Live Demo: https://dynamicheart.github.io/llm-eval-viewer/ Features Multi-format support: Evalscope Predictions / Reviews evaluation results Directory browsing: Select a directory to auto-scan its structure, quickly switch between experiments and datasets (Chrome/Edge) Statistics & distribution: Token histogram, result/finish-reason distribution, per-dataset accuracy — all interactive with click-to-filter Reasoning support: Displays reasoning content marked as [R], view Text and Reasoning separately Dark mode: Light / Dark / Auto theme with system preference detection cURL export: Generate cURL commands from request details for quick API replay i18n: English and Chinese language support Example Files (evalscope) You can use the following example files for local or online experience: Predictions math_500_level_1_predictions.jsonl Reviews math_500_level_1_reviews.jsonl Predictions (with reasoning) humaneval_predictions_with_reasoning.jsonl Reviews (with reasoning) humaneval_reviews_with_reasoning.jsonl Screenshots Reviews View (Dark Mode) Predictions View (Light Mode) Development cd llm-eval-viewer npm install npm run dev Build npm run build Built with AI This project was developed entirely through AI-assisted programming using Claude Opus 4.6. Fr","listed_at":1789166924,"md_updated_at":1789166924},{"full_name":"bluzky/mau","url":"/r/bluzky/mau","github":"https://github.com/bluzky/mau","title":"mau","tagline":"Template engine for Prana workflow which support subset of Liquid ","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Elixir","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Mau, a Liquid-like template engine for Elixir workflows, whose README says \"95% of Mau is written by Claude code under my supervisor\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Mau Template Engine Mau is a powerful Elixir template engine, designed for workflow automation and dynamic content generation, Mau provides a Liquid-like syntax with advanced features including comprehensive filter support, complex expressions, and workflow integration capabilities. 95% of Mau is written by Claude code under my supervisor Features 🔥 High Performance: focus on most vital features 🎯 Feature Rich: 95% implementation coverage of documented features 🔗 Filter Chaining: Full support for complex multi-filter expressions 🎨 Whitespace Control: Precise control over output formatting 🧮 40+ Filters: Comprehensive string, collection, math, and number filters 🌊 Liquid-like Syntax: Familiar template syntax for easy adoption Installation Add mau to your list of dependencies in mix.exs: def deps do [ {:mau, \"~> 0.6\"} ] end Quick Start Basic Usage # Simple template rendering template = \"Hello {{ user.name }}!\" context = %{\"user\" => %{\"name\" => \"Alice\"}} {:ok, result} = Mau.render(template, context) # result: \"Hello Alice!\" Advanced Examples # Complex template with filters and conditionals template = \"\"\" <h1>Welcome {{ user.name | capitalize }}!</h1> {% if user.admin %} <div class=\"admin-panel\">Admin Dashboard</div> {% endif %} <ul> {% for item in items | slice(0, 5) %} <li>{{ item.name | upper_case }}</li> {% endfor %} </ul> <p>Total items: {{ items | length }}</p> \"\"\" context = %{ \"user\" => %{\"name\" ","listed_at":1789166922,"md_updated_at":1789166922},{"full_name":"AdrianEddy/burn-tracing-backend","url":"/r/AdrianEddy/burn-tracing-backend","github":"https://github.com/AdrianEddy/burn-tracing-backend","title":"burn-tracing-backend","tagline":"Tracing backend and trace explorer for the Burn deep learning framework","source":"trawl","tier":"found","license":"Unlicense","stars":8,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tracing backend and trace explorer for the Burn deep learning framework whose README says \"Most of the code was generated by Claude Opus 4.6\". It carries the Unlicense license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Tracing backend for Burn Deep Learning Framework A backend decorator for Burn that traces every tensor operation with CPU timing, detects GPU synchronization points, captures fusion blocks, and generates interactive HTML visualizations for performance analysis. Features Operation tracing — records every float, int, bool, module, and activation operation with precise CPU-side timing Fusion block detection — identifies ElementWise, Matmul, Reduce, and ReduceBroadcasted fusions, including which ops were fused and fallback ops that ran as separate kernels Kernel tracing — captures compiled kernel sources from CubeCL compilation, attached to fusion events for inspection (opt-in via trace-data feature) Tensor shape tracking — records input and output shapes for every operation and fusion block Memory allocation tracking — estimates memory allocations in bytes with per-operation timing Optional tensor data capture — previews up to N tensor values per operation (configurable, opt-in via trace-data feature) Call site tracking — captures file:line:col of the calling code via backtrace (opt-in via trace-caller feature, not available for fused ops) Custom markers and spans — insert markers and timed spans with debug strings, binary data, and tensor attachments GPU sync point detection — explicit (sync()) and implicit (into_data()) synchronization points are highlighted on the timeline. into_data forces a GPU flush and readback, shown clearly on the graph Statistics — total wall time, ope","listed_at":1789166921,"md_updated_at":1789166921},{"full_name":"YazaiHu/MemberHub","url":"/r/YazaiHu/MemberHub","github":"https://github.com/YazaiHu/MemberHub","title":"MemberHub","tagline":"A 100% AI-generated Vibe Coding Member Manager built with Claude Code.","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a chain-store membership management system in Go and Vue, described as \"A 100% AI-generated Vibe Coding Member Manager built with Claude Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"MemberHub - 连锁会员管理系统 一个会员，通行全店 | One Member, All Stores 专为连锁门店打造的统一会员管理系统 功能特性 • 快速开始 • 技术栈 • AI 开发说明 🤖 AI 辅助开发说明 ⚡ 本项目是一个完整的 Vibe Coding 示例项目 使用 Claude Code (claude-sonnet-4.5 模型) 从零到一完整生成，展示了 AI 辅助开发的完整流程和最佳实践。 🎯 为什么创建这个项目？ 这个项目旨在展示： 📚 AI 辅助全栈开发的完整流程 从需求分析到系统设计 从后端 API 到前端界面 从数据库设计到部署方案 🚀 Claude Code 的实际应用能力 复杂业务逻辑实现（积分系统、充值营销） 高并发场景处理（库存控制、分布式锁） 完整的前后端集成 💡 Vibe Coding 开发模式 快速迭代、敏捷开发 AI 与开发者协作 高质量代码生成 📊 开发过程统计 指标 数据 开发模型 Claude Sonnet 4.5 开发工具 Claude Code CLI 总开发时间 ~8 小时（传统开发需 3-4 周） 代码行数 后端 ~15,000 行 / 前端 ~8,000 行 API 端点 35 个 RESTful 接口 前端页面 8 个功能模块 数据库表 20+ 张表，完整的关系设计 测试覆盖 包含完整的 API 测试脚本 文档产出 15+ 份技术文档 🎨 AI 生成内容 ✅ 100% 后端代码：Go + Gin 框架，DDD 架构 ✅ 100% 前端代码：Vue 3 + Element Plus，完整管理后台 ✅ 100% 数据库设计：MySQL 表结构、索引、迁移脚本 ✅ 100% 项目文档：API 文档、部署文档、测试文档 ✅ 配置文件：Docker、Nginx、CI/CD 配置 ✅ 测试脚本：自动化测试、集成测试 🏆 技术亮点 通过 AI 辅助开发，本项目实现了： 高并发场景：积分兑换无超卖（验证 10000+ 并发） 数据一致性：充值事务 100% 原子性保证 分布式架构：Redis 缓存、RabbitMQ 消息队列 安全性：JWT 认证、RBAC 权限、操作审计 性能优化：乐观锁、分布式锁、缓存策略 详细开发笔记： 查看 AI_DEVELOPMENT_NOTES.md 🎨 界面预览 管理后台界面 仪表板 会员管理 门店管理 积分管理 充值管理 优惠券管理 促销管理 💡 提示：所有界面均为实际运行截图，展示完整的管理后台功能。 📊 统计：7 张高清截图，总计 1.3MB，已优化压缩。 📖 项目简介 MemberHub（会员通）是一个专为连锁门店设计的统一会员管理系统，支持跨店积分共享、充值营销、优惠券推送等核心功能。 核心价值 🏪 统一会员体系：用户在所有门店共享积分和余额 🎯 分店自主管理：每个分店可独立设置兑换规则和活动 💰 营销闭环：充值送、优惠券、特价商品等完整营销工具 📱 微信小程序：完整的会员端小程序支持 🎨 管理后台：现代化、响应式的 Web 管理界面 ✨ 功能特性 会员管理 微信小程序一键登录 会员等级体系（普通/白银/黄金/VIP） 会员资产概览（积分/余额/优惠券） 手动调整积分/余额（带备注和审计） 积分系统 ⭐ 消费自动积分（可配置积分比例） 跨店积分统一管理 分店独立设置兑换规则 积分兑换商品/服务 兑换码核销系统 每日限兑控制 高并发库存控制（Redis + 乐观锁） 充值营销 ⭐ 充值送活动配置（如：充 200 送 5","listed_at":1789166900,"md_updated_at":1789166900},{"full_name":"heznpc/skillBridge","url":"/r/heznpc/skillBridge","github":"https://github.com/heznpc/skillBridge","title":"skillBridge","tagline":"Chrome extension for translating supported AI courses into 32 languages with local study tools and exam-safe translation","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["extension"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":690,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a browser extension that translates free AI courses into 32 languages, which its README says \"was developed using Claude Code as an AI pair-programming partner\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"SkillBridge — AI Course Translator Source checkpoint: v4.2.0 — released on GitHub only, not designated for Chrome Web Store upload. The live CWS version remains v1.0.1; final packaging and dashboard work are deferred until the current development cycle is complete. Available in multiple languages at the project landing page. Translate the free AI courses at anthropic.skilljar.com into your language — instantly. Break the language barrier on these free AI courses. 32 languages supported. The CWS edition combines course-page translation with local flashcards, bookmarks, progress, reading, and export tools. It runs on anthropic.skilljar.com, detected Skilljar-hosted AI courses, Claude Academy course routes at academy.claude.com/courses, and Claude tutorial pages at claude.com/resources/tutorials; non-AI Skilljar tenants are paused automatically. Version boundary: the Chrome Web Store still serves legacy v1.0.1, which includes the Puter-backed Gemini/Claude path and YouTube host permission. References below to the “CWS edition” describe the intended package boundary, not an assigned upload candidate. Publication is paused, and the final CWS version will be assigned after ongoing development; see the version-split Privacy Policy. Automated browser fixtures cover Academy routing and the Tutor transport, but this checkpoint does not claim an authenticated live Academy post-submit DOM capture or a final signed-in Puter round trip; both remain final CWS gates. Install · Features · Rep","listed_at":1789166899,"md_updated_at":1789166899},{"full_name":"xmonkey/MathAcademyLogAnalyzer","url":"/r/xmonkey/MathAcademyLogAnalyzer","github":"https://github.com/xmonkey/MathAcademyLogAnalyzer","title":"MathAcademyLogAnalyzer","tagline":"A Python tool to parse, analyze, and visualize learning progress from MathAcademy.com's Activity Log PDF reports.","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Python tool that charts Math Academy activity log PDFs, which its README describes as \"built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"MathAcademyLogAnalyzer MathAcademyLogAnalyzer is a Python tool for analyzing course progress data from mathacademy.com's PDF activity logs, built with Claude Code. Generate charts and statistics from your Math Academy activity logs to track learning progress and identify patterns. Inspired by rng.eth 中文文档 📊 Chart Gallery Learning Dashboard Learning Progress Analysis Cumulative XP Trend Daily XP Distribution Multi-Level Statistics Task Type Distribution Efficiency Trend Daily XP Histogram Weekday Performance Quick Start 1. Install # Clone and install git clone https://github.com/xmonkey/MathAcademyLogAnalyzer.git cd MathAcademyLogAnalyzer pip install -e . Windows users: pip install -e . 2. Download Activity Log PDF Log in to mathacademy.com with your parent/supervisor account Click the student's settings icon (⚙️) next to their name Select \"Documentation\" from the menu Under \"Activity Log\", click \"Request...\" Important: Set time frame to include your entire learning period (recommended: from first day) Click \"Preview\" → Download the PDF file 3. Analyze mathacademy-analyzer generate-all activity_log.pdf -o output_folder Output Formats Interactive HTML (default): Interactive charts with zoom, hover, and tooltips Static PNG: Static image files suitable for documents and sharing Excel: Structured data with course details and activity logs JSON: Raw data for custom analysis More Useful Commands # Get PDF information mathacademy-analyzer info activity_log.pdf # Extract text or tabl","listed_at":1789166894,"md_updated_at":1789166894},{"full_name":"rallisf1/greek-postal-code-db","url":"/r/rallisf1/greek-postal-code-db","github":"https://github.com/rallisf1/greek-postal-code-db","title":"greek-postal-code-db","tagline":"Embed a postcode database into your (offline) application. Regions, municipalities, communities, and partial street coverage included.","source":"trawl","tier":"found","license":"MIT","stars":15,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Multi-language clients embedding a Greek postal code database for offline lookup and address validation; \"These packages have been almost entirely AI generated.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Greek Postal Code DB clients Language clients for the Greek Postal Code DB. Each package embeds the same read-only library.sqlite asset and is versioned and published independently. Features 1328 postal codes with hierarchy information (region, municipality, community, etc) partial street database (which street and numbers belong to each postal code) official government codes from Kallikratis (WIP) sqlite database included for offline use list / get / search / validateAddress functions Packages typescript: Node.js and Bun client, published as @rallisf1/greek-postal-code-db. php: PHP 8.1+ Composer client, published from this repository root as rallisf1/greek-postal-code-db. go: Go client, imported as github.com/rallisf1/greek-postal-code-db/go. python: Python client; install from this monorepo with pip's #subdirectory=python URL. rust: Rust crate; use Cargo's Git subdirectory = \"rust\" support until crates.io publication. csharp: .NET 10 client, packaged for NuGet as GreekPostalCodeDb. java: Java 21 client, packaged for Maven Central as io.github.rallisf1:greek-postal-code-db. dart: native Dart client, published as greek_postal_code_db. flutter: Flutter asset-bundle adapter, published as greek_postal_code_db_flutter. swift: native Swift Package Manager client for Linux and Apple platforms. models: private TypeScript types used internally; it is not published. demo: internal Cloudflare Pages/D1 demo application; it is not published. Contributing Packages These packages have been","listed_at":1789166892,"md_updated_at":1789166892},{"full_name":"Adudeguyman/comfyui_fantastic-loras","url":"/r/Adudeguyman/comfyui_fantastic-loras","github":"https://github.com/Adudeguyman/comfyui_fantastic-loras","title":"comfyui_fantastic-loras","tagline":"Vibe coded nodes based on smarter people to make their interface the way I want it.","source":"trawl","tier":"found","license":"MIT","stars":17,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A ComfyUI LoRA toolkit with a 12-slot loader and XY test bench, described by its owner as \"Vibe coded nodes based on smarter people\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Fantastic Loras 2.0 A LoRA toolkit for ComfyUI: pick which LoRA subfolders a node draws from so you're only ever searching the models you're working with, stack up to 12 LoRAs across several models from one node, and run a proper XY LoRA test bench in a single queue. The Fantastic Lora Loader with 12 lora slots and presets Now with Themes! Filter which subdirectories are used when searching for loras Search for a Lora from only your selected folders. Save your favorites for quick access Adjust loras strength for each connected model Pick your Loras and the strengths to try, add a control image, and queue once — every combination generates on its own Effortlessly generate multiple iterations to compare Loras and their strengths. Rearrange the output as you see fit Choose between modern and classic grid outputs What it does One loader instead of a chain of them. Twelve LoRA slots in a single node, feeding up to five separate MODEL outputs. Each LoRA can go to all five models, or just one, or to different models at different strengths — so a two-pass workflow where the refiner needs your detail LoRA at 0.4 and the base pass wants it at 1.0 is one node, not two loaders and a mental note. Slot order is apply order and you drag slots to change it. A footer prints exactly what each model output will receive, in order, so you can confirm the wiring matches your intent without tracing wires. Alongside that: named presets that save and reload your whole stack (or merge into it without ","listed_at":1789166891,"md_updated_at":1789166891},{"full_name":"GZT2023/ComfyUI-Qwen3.5-Low-VRAM-GPU","url":"/r/GZT2023/ComfyUI-Qwen3.5-Low-VRAM-GPU","github":"https://github.com/GZT2023/ComfyUI-Qwen3.5-Low-VRAM-GPU","title":"ComfyUI-Qwen3.5-Low-VRAM-GPU","tagline":"ComfyUI plugin for running Qwen3.5 models locally with  low VRAM optimization ","source":"trawl","tier":"found","license":"Apache-2.0","stars":12,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A ComfyUI plugin for running Qwen3.5 image captioning on low-VRAM GPUs; \"This plugin is entirely AI-generated and has been successfully run.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"充电支持：🔗 B站：赛博画师GZT ComfyUI-Qwen3.5-Low-VRAM-GPU 针对12G低显存GPU、32G低内存深度优化, 专为ComfyUI设计的Qwen3.5图像描述插件。 第一次运行，自动从国内ModelScope下载模型，完全阻断HuggingFace连接。下载模型之后，即可拔掉网线运行，可完全离线，网络环境友好。 支持量化加载与智能显存管理。支持 Qwen3.5-2B/4B/9B/27B/35B-A3B 等模型。 本插件使用纯AI生成，已成功运行。 以下内容也使用AI生成。 This plugin is entirely AI-generated and has been successfully run. The following content is also generated using AI. ✨ 核心特性 🚀 纯国内加速：默认从ModelScope下载模型，彻底规避HuggingFace连接问题，支持完全离线运行。 💾 极致显存管理： 自动切换卸载：切换模型时，自动卸载旧模型，杜绝资源残留。 推理后可选卸载：生成描述后（默认开启），立即从显存/内存中卸载模型，为后续工作流（如生图）腾出全部空间。 📊 低显存量化：支持 4bit 和 8bit 量化，加载2B/4B模型仅需6GB显存。 🖥️ 优雅的文本显示：自带前端节点，直接显示生成的描述，支持框选复制，并完美适配深色/浅色主题。 🔌 即插即用：完全集成ComfyUI节点系统，无需复杂配置。 📦 安装 方法一：直接下载（推荐） 进入ComfyUI的 custom_nodes 目录。 克隆本仓库： cd custom_nodes git clone https://github.com/GZT2023/ComfyUI-Qwen3.5-Low-VRAM-GPU.git 安装依赖： pip install modelscope bitsandbytes opencv-python accelerate transformers 注意：bitsandbytes 在Windows上可能需要预编译版本，如遇安装问题，可暂时不使用量化（选择 none）。 方法二：手动安装 下载本仓库的ZIP压缩包，解压至 custom_nodes/ComfyUI-Qwen3.5-Low-VRAM-GPU/。 在插件目录下，使用ComfyUI的Python环境运行： pip install -r requirements.txt （请确保您已创建包含上述依赖的 requirements.txt 文件） 便携包环境的话，在便携包目录ComfyUI_windows_portable下运行： .\\python_embeded\\python.exe -m pip install -r .\\ComfyUI\\custom_nodes\\ComfyUI-Qwen3.5-Low-VRAM-GPU\\requirements.txt 🚀 快速开始 重启ComfyUI，在节点菜单中找到分类 ComfyUI-Qwen3.5-Low-VRAM-GPU。 添加以下三个节点，并按下图连接： LoadQwen35Model (Low VRAM)：选择模型（如 Qwen3.5-2B）和量化等级。 Qwen35Caption (Low VRAM)：连接模型和待处理的图像。 ShowCaptionText (Low VRAM)：连接生成的文本，并直接显示结果。 执行工作流。首次使用且未勾选 local_files","listed_at":1789166888,"md_updated_at":1789166888},{"full_name":"andrew-savage/ha-ptv","url":"/r/andrew-savage/ha-ptv","github":"https://github.com/andrew-savage/ha-ptv","title":"ha-ptv","tagline":"PTV integration for Home Assistant. Can pull all train data to basically replicate the PTV app.","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Home Assistant integration for Melbourne PTV trains and trams, whose README says \"This project was written entirely by AI.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"AI Disclosure: This project was written entirely by AI. It has been tested on multiple HA installs. PTV – Home Assistant Integration A Home Assistant custom integration for Public Transport Victoria (PTV) metro trains and trams. Provides real-time delays, cancellations, service alerts, vehicle positions, and scheduled departure times for any Melbourne Metro station or tram stop. Features Trains and trams — all 16 Metro lines and all 24 tram routes, each added as its own entry Real-time trip updates — delays, cancellations, schedule relationships Service alerts — published PTV disruption messages (signal faults, planned works, etc.) Vehicle positions — count of services currently running, plus worst occupancy level (trains only) Next N departures per station per direction — realtime where available, static timetable fallback otherwise Express services flagged — departures running express are marked and labelled in the displayed time Overnight works handled sensibly — planned nightly works are listed with their actual hours instead of reporting the line as disrupted all day Adaptive polling — polls fast during commute peaks and slowly overnight and at weekends, cutting API usage by around 64% Problem level sensor — none / delays / alert / cancellations / major for easy automation branching Problem summary sensor — TTS-ready plain English sentence Multiple stations — monitor any number of stations, each with configurable directions and departure count Static timetable fallback —","listed_at":1789166887,"md_updated_at":1789166887},{"full_name":"petripaavola/IntuneDeviceDetailsGUI","url":"/r/petripaavola/IntuneDeviceDetailsGUI","github":"https://github.com/petripaavola/IntuneDeviceDetailsGUI","title":"IntuneDeviceDetailsGUI","tagline":"Intune Device Details GUI HTML v4.2 (New Updated!)","source":"trawl","tier":"found","license":"MIT","stars":417,"language":"PowerShell","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a PowerShell tool that builds HTML reports on Intune-managed devices, whose README says the \"Settings Catalog parsing logic was 100% AI-generated\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🚀 Intune Device Details GUI HTML - The Ultimate Report Tool Version 4.23 - The Next Generation Completely rebuilt from the ground up with AI-powered architecture 🐛 Report Issues | 💡 Feature Requests ⚡ Quick Start Download IntuneDeviceDetailsGUI.ps1 and unblock (Properties → Unblock) Install module: Install-Module Microsoft.Graph.Authentication -Scope CurrentUser Allow scripts: Set-ExecutionPolicy RemoteSigned -Scope CurrentUser Run: .\\IntuneDeviceDetailsGUI.ps1 Approve permissions when prompted (first run only) - See required permissions Search device by name, serial, or user (use ? for built-in filters like \"synced in last 24h\") Select Extended Report (default) for full features (conflict detection, scripts) HTML report opens automatically in browser 🎯 What This Tool Does The \"Resultant Set of Policy\" for Intune - Understand WHAT should be configured and WHY on device based on all assigned apps, policies, scripts, groups and filters. Complete Device Intelligence (v4.x) Device: Hardware, OS, storage, network, sync, Autopilot, enrollment type, MDE status Users: Primary + logged-in users, group memberships with dynamic rules Apps: All Assignments for Device with install status, Win32 scripts (install/uninstall/detection/requirement), filters, application information Policies: All Assignments for Device with Complete settings for few policy types (e.g. Settings Catalog, OMA-URI), conflict detection ⭐ Scripts: All Assignments for Device Remediation/platform/shell scripts (cle","listed_at":1789166884,"md_updated_at":1789166884},{"full_name":"Isobastian/Notte-Accessibility","url":"/r/Isobastian/Notte-Accessibility","github":"https://github.com/Isobastian/Notte-Accessibility","title":"Notte-Accessibility","tagline":"Free, open-source browser extension that adapts any website for low vision — high-contrast dark mode plus a per-site accessibility toolkit. Chrome, Firefox, Safari. ","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["extension"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a low-vision accessibility and dark mode browser extension, whose README says its \"source code is written by an AI assistant working from my specifications\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🌙 Notte — Accessibility & Dark Mode From dark mode to full accessibility. A free, open-source browser extension that adapts any website in real time to be easier to see and read — built as an accessibility tool for the low-vision community, and useful to anyone who finds bright pages hard on the eyes. Free forever: no ads, no donations, no tracking, no data collection. One codebase, three browsers — Chrome, Firefox and Safari (iPhone, iPad and Mac). Notte is a user-side tool: it adapts pages for you, in your own browser. It is the opposite of a site-owner \"accessibility overlay\" — it never changes the site for anyone else, and never makes claims about a site's compliance. Install Notte is free, on all three stores. Browser Get it Chrome, Edge, Brave Install from the Chrome Web Store Firefox (desktop + Android) Install from Firefox Add-ons Safari (iPhone, iPad, Mac, Vision) Get it on the App Store Prefer to build it yourself, or want the latest development version? See Building from source below. What Notte does today Dark mode A high-contrast dark-mode engine that darkens overly bright websites by remapping colors, never by inverting them, so contrast stays high and the page keeps its identity: 🌑 Light backgrounds become dark, light text becomes bright, and accent colors (buttons, badges, selected rows) stay distinct — contrast averages ~11:1 in testing 🎯 Leaves already-dark sites alone — it detects a site that already ships its own dark theme and stays out of the way ","listed_at":1789166883,"md_updated_at":1789166883},{"full_name":"yuchanns/ishiku.nvim","url":"/r/yuchanns/ishiku.nvim","github":"https://github.com/yuchanns/ishiku.nvim","title":"ishiku.nvim","tagline":"Neovim treesitter parser manager and runtime integration for vim.treesitter.","source":"trawl","tier":"found","license":"Apache-2.0","stars":3,"language":"Lua","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Neovim plugin that installs and manages Tree-sitter parsers and textobjects, whose README cautions that it is \"entirely AI-generated\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"ishiku.nvim Neovim 0.11 and 0.12 Treesitter parser manager and runtime integration for vim.treesitter. ishiku.nvim manages parser installation, activates registry-provided runtime files, auto-starts Treesitter for supported buffers, and provides a textobjects module compatible with the common nvim-treesitter workflow. Use it together with ishiku-registry. CautionThis repository is entirely AI-generated. Features Install and update Treesitter parsers. Activate registry runtime assets such as queries/*/*.scm. Auto-start vim.treesitter for file-backed buffers. Provide textobjects support on top of vim.treesitter: select move swap repeatable_move incremental_selection lsp_interop Setup require(\"ishiku\").setup { ensure_installed = { \"lua\", \"vim\", \"python\" }, auto_install = true, sync_install = false, registries = { \"github:yuchanns/ishiku-registry\" }, } Textobjects ishiku.nvim can expose a textobjects configuration model close to nvim-treesitter-textobjects. require(\"ishiku\").setup { ensure_installed = { \"lua\", \"vim\", \"go\", \"rust\", \"python\", \"typescript\" }, auto_install = true, textobjects = { select = { enable = true, lookahead = true, keymaps = { [\"af\"] = \"@function.outer\", [\"if\"] = \"@function.inner\", [\"ac\"] = \"@class.outer\", [\"ic\"] = \"@class.inner\", [\"aa\"] = \"@parameter.outer\", [\"ia\"] = \"@parameter.inner\", }, }, move = { enable = true, set_jumps = true, goto_next_start = { [\"]m\"] = \"@function.outer\", [\"]]\"] = \"@class.outer\", }, goto_next_end = { [\"]M\"] = \"@function.outer\", [\"][","listed_at":1789166881,"md_updated_at":1789166881},{"full_name":"thieso2/MoremaidApp","url":"/r/thieso2/MoremaidApp","github":"https://github.com/thieso2/MoremaidApp","title":"MoremaidApp","tagline":"Native macOS Markdown viewer with Mermaid diagram support","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a native macOS Markdown viewer with Mermaid diagram support, whose README says \"This project is 100% AI-generated.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Moremaid A native macOS Markdown viewer with full Mermaid diagram support. Browse markdown files with beautifully rendered diagrams, syntax-highlighted code, full-text search, encrypted archives, and 10 color themes. This project is 100% AI-generated. Every line of code, feature, architectural decision, and this README were created by Claude (Anthropic's AI). The human provided requirements and feedback; Claude handled all programming, debugging, and problem-solving. Features Rendering Mermaid diagrams -- flowcharts, sequence diagrams, Gantt charts, state diagrams, ER diagrams, and more Syntax highlighting for 20+ languages via Prism.js Fullscreen diagrams -- click to open any Mermaid diagram in its own window Live reload -- file changes are detected and re-rendered automatically Themes & Typography 10 color themes -- Light, Dark, GitHub, GitHub Dark, Dracula, Nord, Solarized Light/Dark, Monokai, One Dark 6 typography styles -- Default, GitHub, LaTeX, Tufte, Medium, Compact Theme selector in Preferences System dark mode detection Search Fuzzy filename search with relevance ranking Full-text content search with context snippets and match highlighting Toggle between modes with TAB Keyboard-driven navigation (/ to focus, arrows to browse, Enter to open) Archive & Sharing Pack markdown projects into .moremaid archives AES-256 encryption -- optional password protection In-memory browsing -- encrypted files are never written to disk Zero-install viewing -- open .moremaid ar","listed_at":1789166880,"md_updated_at":1789166880},{"full_name":"davideconsonni/authglow","url":"/r/davideconsonni/authglow","github":"https://github.com/davideconsonni/authglow","title":"authglow","tagline":"OAuth 2.0 & OpenID Connect authorization server","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":634,"vulnerable":16},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted OAuth 2.0 and OpenID Connect server that stores everything as files, whose README says \"100% AI-generated.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"AuthGlow Self-hosted OAuth 2.0 / OpenID Connect authorization server with no database. Files in, JWTs out — swap storage backends with one environment variable. What is AuthGlow? AuthGlow is a self-hosted OAuth 2.0 / OpenID Connect authorization server, user directory, and admin console — with no database to run, patch, or back up. Users, sessions, tokens, and OAuth2 clients are stored as files through an fsspec abstraction, so the exact same code runs on your laptop, a VPS, or against an S3 bucket. Change STORAGE_BACKEND from file to s3, gcs, or abfs and your data — including the JWT signing keys — moves with it. No migrations, no schema, no code changes. 🎯 Key Features 🔐 OAuth 2.0 / OIDC Authorization Server — Full implementation with PKCE, JWKS auto-rotation, DPoP (RFC 9449), Device Authorization Grant (RFC 8628), and refresh token theft detection 🛡️ Multi-Factor Authentication — TOTP, backup codes, trusted devices, seamlessly integrated into OAuth2 flows 🔑 Passkeys / WebAuthn / FIDO2 — Passwordless authentication with biometrics (Touch ID, Windows Hello) and security keys (YubiKey) 🌍 Identity Federation — Login via CIE, SPID, Google, Microsoft/Entra ID, Apple, Keycloak, Auth0, Okta — any OIDC provider works out of the box ☁️ Serverless & Simple — No database required, deploy anywhere in 30 seconds. Storage backend swappable via environment variable (file, S3, GCS, Azure Blob) ✨ Features Authentication & protocols OAuth 2.0 & OpenID Connect — Authorization","listed_at":1789166878,"md_updated_at":1789166878},{"full_name":"vahabahmadvand/VibeNotes","url":"/r/vahabahmadvand/VibeNotes","github":"https://github.com/vahabahmadvand/VibeNotes","title":"VibeNotes","tagline":"A lightweight, developer-focused Windows desktop application inspired by Windows Sticky Notes, with native Markdown editing, syntax-highlighted collapsible code blocks, SQLite persistence, and native Windows Startup / System Tray integration. Vibe-coded with @Google-Antigravity","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Tauri sticky notes desktop app for Windows with Markdown and code blocks, which its owner says was \"Vibe-coded with @Google-Antigravity\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"VibeNotes 📝⚡ A lightweight, developer-focused Windows desktop application inspired by Windows Sticky Notes, with native Markdown editing, syntax-highlighted collapsible code blocks, SQLite persistence, and native Windows Startup / System Tray integration. Built with Tauri v2 + Rust + React 19 + TypeScript + Tailwind CSS (Under 40 MB idle RAM, zero Chromium bloat). 📸 Screenshots 🪟 Floating Desktop Sticky Notes Multi-theme floating notes with interactive checklists, VS Code Dark+ code blocks, and markdown tables: 🗄️ Central Notes Hub Search notes, filter by color, toggle card sizes, and manage your entire workspace: 🔄 Seamless In-App Updates Check for GitHub releases, view changelogs, download signed updates, and restart with one click: ✨ Features 🪟 Floating Frameless Sticky Notes: Custom drag regions, real-time coordinate & size persistence, and per-note \"Always on Top\" pin. 🚀 Windows Startup & System Tray: Run in background with a native tray context menu (New Note, Notes Hub, Show/Hide All, Startup Toggle, Exit). 🎨 7 Color Themes: Classic Yellow, Mint Green, Pastel Pink, Lavender Purple, Sky Blue, Charcoal Dark, and Slate Grey. 💻 Developer Markdown & Code Blocks: Multi-language syntax highlighting with VS Code Dark+ styling. Collapsible / Expandable long code snippets with line count badges. One-click \"Copy Code\" button. Interactive checklists with nested task lists (Tab / Shift+Tab). Tables, blockquotes, links, and drag-and-drop image support. 🗄️ Centr","listed_at":1789166874,"md_updated_at":1789166874},{"full_name":"WillPresley/scribenode-podcast-transcription","url":"/r/WillPresley/scribenode-podcast-transcription","github":"https://github.com/WillPresley/scribenode-podcast-transcription","title":"scribenode-podcast-transcription","tagline":"ScribeNode is a full-stack, high-throughput AI audio transcription and speech intelligence web application.","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":333,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted podcast and audio transcription web app, whose README says it \"was built using AI-assisted pair programming (\"vibecoded\")\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"ScribeNode — AI Speech & Transcript Engine ScribeNode is a full-stack, high-throughput AI audio transcription and speech intelligence web application. Powered by Google's specialized Gemini Flash and Audio AI model suite (gemini-3.8-flash, gemini-3.7-flash, gemini-3.6-flash, and gemini-3.5-flash), ScribeNode transforms raw podcast recordings, meeting audio, interviews, and voice notes into polished clean-verbatim transcripts, structured chapters, executive summaries, and actionable key takeaways. Note on Development: This project was built using AI-assisted pair programming (\"vibecoded\") and then manually audited, refined, and tested for code quality, type safety, and container security. Key Features 🎧 Broad Format & Configurable Uploads: Transcribe MP3, WAV, M4A, OGG, and FLAC audio files with configurable file size limits (MAX_UPLOAD_SIZE_MB, defaults to 100MB). 📻 Podcast RSS Feed & Remote URL Ingestion: Directly transcribe episodes from public podcast RSS feeds or direct audio URLs with live feed inspection and episode selection. 📖 Custom Vocabulary & Technical Glossary: Inject domain-specific terminology, brand names, acronyms, and proper nouns into transcription prompts to maximize recognition accuracy. ✏️ Inline Quick Edit & Global Speaker Renaming: In-place transcript editing via markdown editor and one-click global speaker alias replacement across entire transcripts. 🎙️ High-Fidelity Audio Transcription: Multimodal transcription with gemini-3.8","listed_at":1789166873,"md_updated_at":1789166873},{"full_name":"AlexTuring010/signal-processing-with-alexturing","url":"/r/AlexTuring010/signal-processing-with-alexturing","github":"https://github.com/AlexTuring010/signal-processing-with-alexturing","title":"signal-processing-with-alexturing","tagline":"Interactive Greek-language study companion for the K21 Communication Systems course (UoA DI, 4th semester). Next.js 15 + MDX + KaTeX + Supabase comment system + gamification (orchard, collectibles).","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["web-app"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an interactive Greek-language study site for a university signal-processing course, which its owner calls an experiment in \"collaborative authoring with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"signal-processing-with-alexturing An interactive study companion for the K21 — Communication Systems course at the University of Athens (Department of Informatics & Telecommunications, 4th semester). Built to help my Greek-speaking classmates understand the material rather than memorize it. Live: signal-processing-with-alexturing.vercel.app The site is in Greek because its audience is. This README is in English for anyone landing on the repo from outside the course. What's on the site Long-form lessons in MDX, structured around the course syllabus. Each section has its own pedagogical plan in plans/ that gets handed to Claude Code one at a time for incremental drafting. Interactive math — KaTeX-rendered equations, <Eq id=\"…\" /> for named formulas you can cross-reference, sliders and Canvas/SVG demos for signal-shaping intuition. Comment system on every section — Supabase-backed, magic-link or Google sign-in. Authors can edit/delete their own posts within 10 minutes; moderator role for the rest. Includes a special is_claude_reply flag for assistant-authored replies (moderator-gated). Practice hub with past exam problems indexed by topic. Formula sheet as an interactive page rather than a PDF. Gamification — an \"orchard\" with trees that produce apples for engagement, plus per-section collectible skins. The most recent commits in the repo's history are working through Phase 5 of the collectibles roll-out. Sibling project — algorithms class This repo also hosts an algor","listed_at":1789166870,"md_updated_at":1789166870},{"full_name":"Alamin-H-M/manim-slides-preview","url":"/r/Alamin-H-M/manim-slides-preview","github":"https://github.com/Alamin-H-M/manim-slides-preview","title":"manim-slides-preview","tagline":"VS Code extension: one-click render → live auto-refreshing preview (browser / VS Code tab / native GUI) + background .pptx export for manim-slides. Offline installable, zero dependencies. Open source, AI-generated.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["extension"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a VS Code extension that live-previews manim-slides presentations, whose README says \"The implementation was written by an AI assistant working under Alamin's direction\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Manim Slides Preview Live preview for manim-slides in VS Code. 50-second demo video Press the run button once. After that, every Ctrl+S re-renders your changes and refreshes the interactive presentation in a browser tab, a VS Code tab, or the native presenter window. A PowerPoint copy is kept up to date in the background. No terminal commands. The extension is a single 28 KB file with zero runtime dependencies and zero telemetry. Everything runs locally: the generated presentation bundles Reveal.js, so it works without internet. New to LaTeX, Manim, or manim-slides? The free offline course Manim Slides Academy teaches the whole stack around this workflow. Contents Requirements Install First presentation The workflow Coming from plain Manim What the extension replaces Presenting How it stays fast Commands Settings Troubleshooting Building from source Credits 1. Requirements One-time setup per machine: Tool Verify with Install Python 3.9+ python --version python.org Manim CE ≥ 0.19 manim --version pip install manim manim-slides ≥ 5.1.10 manim-slides --version pip install -U \"manim-slides[pyside6]\" LaTeX (MiKTeX or TeX Live) latex --version Only if your scenes use MathTex / Tex FFmpeg is not required. Manim CE ≥ 0.19 encodes video through its bundled PyAV library. The ffmpegPath setting exists only for older Manim versions and plugins that call an ffmpeg binary directly. 2. Install You need one file: manim-slides-preview-1.7.4.vsix (from Releases). No marketplace, no internet. P","listed_at":1789166869,"md_updated_at":1789166869},{"full_name":"yuansheneh/musicmeta","url":"/r/yuansheneh/musicmeta","github":"https://github.com/yuansheneh/musicmeta","title":"musicmeta","tagline":"纯ai编写 别喷我呀","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a cross-platform music metadata tag editor for Windows and Android, whose README says \"This project is entirely AI-generated.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🎵 MusicMeta - 跨平台音乐元数据编辑器 一个功能强大、界面华丽的跨平台音乐元数据编辑工具，专为 Windows 和 Android 优化，同时支持 macOS 和 Linux。 ✨ 主要特性 🎨 华丽的 GUI 界面 多主题支持: 内置暗黑、明亮、海洋、日落、森林 5 种精美主题 实时预览: 封面艺术和元数据实时显示 现代化设计: 基于 CustomTkinter 的流畅界面 高 DPI 适配: 完美支持 4K 和高分辨率屏幕 📀 多种音频格式支持 MP3 - ID3v1, ID3v2.3, ID3v2.4 FLAC - Vorbis Comments OGG Vorbis/Opus - Vorbis Comments M4A/AAC - iTunes MP4 标签 WMA - ASF 标签 WAV - INFO chunk AIFF - ID3 标签 APE - APEv2 标签 🖥️ Windows 特殊优化 ✅ 文件资源管理器集成 ✅ 右键菜单快速访问 ✅ 任务栏进度显示 ✅ Windows 11 暗色标题栏 ✅ 高 DPI 自动缩放 ✅ 系统主题自动检测 📱 Android 特殊优化 ✅ Material You 动态取色 ✅ 存储访问框架 (SAF) ✅ 后台元数据扫描 ✅ 触摸优化界面 ✅ 移动网络优化 ✅ 媒体扫描器集成 🛠️ 强大的编辑功能 基本元数据: 标题、艺术家、专辑、年份等 高级元数据: 作曲家、指挥家、混音师、出版商等 媒体信息: BPM、情绪、歌词、评级 封面艺术: 上传、查看、移除专辑封面 批量编辑: 同时处理多个文件 查找替换: 快速批量修改元数据 从文件名标记: 自动从文件名提取信息 🚀 快速开始 安装依赖 pip install -r requirements.txt 运行应用 Windows: run.bat Linux/macOS: chmod +x run.sh ./run.sh 直接运行: python run.py 命令行模式 # 显示文件信息 python run.py song.mp3 --info # 启动 GUI python run.py --gui 📖 使用指南 基本使用 打开文件: 点击\"📁 打开文件\"按钮选择音频文件 编辑元数据: 在右侧编辑器中修改元数据字段 更换封面: 点击封面区域上传新封面图片 保存: 点击\"💾 保存\"按钮保存更改 批量编辑 点击\"📂 打开文件夹\"加载整个音乐库 选择多个文件 使用批量编辑功能统一修改元数据 支持查找替换、复制元数据等操作 主题切换 在左侧边栏的主题下拉菜单中选择喜欢的主题： 🌑 暗黑模式 - 护眼舒适 ☀️ 明亮模式 - 清晰明快 🌊 海洋 - 深邃蓝色 🌅 日落 - 温暖橙色 🌲 森林 - 自然绿色 📁 项目结构 musicmeta/ ├── src/ │ ├── core/ # 核心元数据处理引擎 │ │ ├── metadata.py # 元数据模型 │ │ ├── audio_loader.py # 音频加载器 │ │ └── tag_mapper.py # 标签映射器 │ ├── gui/ # 图形用户界面 │ │ ├── main_window.py # 主窗口 │ │ ├── components.py # UI 组件 │ │ └── theme_manager.py # 主题管理器 │ ├── platforms/ # 平台特定优化 │ │ ├── windows_optimizer.p","listed_at":1789166867,"md_updated_at":1789166867},{"full_name":"shazow/go-whatchanged","url":"/r/shazow/go-whatchanged","github":"https://github.com/shazow/go-whatchanged","title":"go-whatchanged","tagline":"Pretty and practical Go library API diffs. 🧐📑","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: go-whatchanged, a CLI that prints API diffs of Go modules between versions; \"It was fairly vibecoded, out of a desperate need to review large pull requests better\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"go-whatchanged What changed in the public API? go-whatchanged prints an API diff of a Go module's API between two versions. Read-only: Happy path is optimized to avoid writing to the filesystem (no temporary directories, git clones, or worktrees). Only go commands are used for comparing remote packages that are not already cached, use --fsreadonly to disable any features that rely on writing the filesystem. Release-aware: All standard Go version tags supported (@latest for latest tagged release; @previous for the release before it, which pairs the last two releases; @v1.2.3 for a specific tag; @HEAD for a ref like HEAD, etc). CI-ready. Markdown for pull requests, JSON for tools, exit codes for gates, and a GitHub Action that posts the diff to the job summary and as a pull request comment. Status: beta. It was fairly vibecoded, out of a desperate need to review large pull requests better, but the readonly constraints add a lot of safety (and efficiency) to how it works. It's quite useful! Example $ go-whatchanged @latest example.com/m/store - func (c *Client) Close() error - func Open(path string) (*Client, error) + func Open(path string, o Options) (*Client, error) + func (c *Client) Ping() error + type Options struct{ Timeout int } example.com/m/util + func (Sizer) Size() int 2 packages changed · 3 incompatible · 2 compatible · would require: MAJOR (v1.4.0 → v2.0.0) Install go install github.com/shazow/go-whatchanged@latest Usage Question Command What do my uncommitted chang","listed_at":1789166865,"md_updated_at":1789166865},{"full_name":"smeshidojoe/Snatchr","url":"/r/smeshidojoe/Snatchr","github":"https://github.com/smeshidojoe/Snatchr","title":"Snatchr","tagline":"Just Download Anything.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["media"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":6,"vulnerable":1},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Snatchr, a Windows system-tray video and audio downloader with a Spotlight-style launcher, whose README notes \"This is a fully vibecoded product\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🇷🇺 Русский · 🇬🇧 English Snatchr Just Download Anything. Snatchr is a fast, lightweight video & audio downloader for Windows that lives in your system tray. Grab media from 1,000+ sites without opening a browser, fighting ads, or touching the command line — then trim, copy, and re-download in a couple of clicks. Website · Guide · Download 🖥️ Lives in your system tray No taskbar window, no clutter — Snatchr sits quietly in the system tray. Left-click the icon to open the app, right-click for quick actions (Paste a link, Open, Exit). While anything is downloading, the tray icon spins, and a green check pops when it's done. ✨ Spotlight The heart of Snatchr. Press Ctrl + E anywhere to summon a Spotlight-style launcher: Paste → done. Drop a link and it downloads in the background at best quality — no dialogs. Several downloads at once, each as its own row with a live progress bar and thumbnail. Unified history of everything you've downloaded (from the window, tray, or Spotlight) — with the source and resolution, newest on top. Trim on the spot — cut a clip with a frame filmstrip range slider and preview, then save it next to the original or copy it to the clipboard. One-click copy of any finished file, straight into a chat or Explorer. The shortcut and dismiss behaviour are customizable in Settings. 🔔 Clipboard toast Turn on clipboard watching and Snatchr keeps an eye out: copy a supported link and a little toast appears offering to grab it. Left-click to download in the ","listed_at":1789166863,"md_updated_at":1789166863},{"full_name":"sirhafizho/charity-community-listing","url":"/r/sirhafizho/charity-community-listing","github":"https://github.com/sirhafizho/charity-community-listing","title":"charity-community-listing","tagline":"A simple Charity Community Listing System - Post, categorize, locate, and claim charity items with admin approval workflow","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a charity donation listing web app with claims and admin approval, whose README says \"This project was built entirely with AI assistance as an experiment in AI-driven software engineering.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🤝 Charity Community Listing System A full-stack charity community listing platform where users can post items for donation, browse by category and location, claim items they need, and track the full donation lifecycle. Includes admin approval, impact dashboards, reporting, and more. 🌐 Live Demo: charity-community-listing.vercel.app 📸 Screenshots Homepage — Browse listings Dark Mode Admin Panel — Approve/Reject listings Create Listing — With image upload & tags Listing Detail — Claim & messaging Dashboard — Impact stats & badges Login Page Mobile Responsive ✨ Features Core 📸 Image Upload — Upload images for listings (JPEG, PNG, WebP, GIF, up to 5MB) 📂 Categories — Organize listings by type (Food, Clothing, Electronics, Furniture, Books, Toys, Health, Other) 🏷️ Tags — Add custom tags for better discoverability 📍 Location — Add location info to help nearby community members 🔍 Search & Filter — Search by keyword, filter by category/location/condition with debounced input 🙋 Claim System — Users can claim items with a message to the donor 🔐 User Authentication — Register/login with email & password (bcrypt hashed) 👨‍💼 Admin Panel — Approve or reject listings before they go public 🌗 Dark Mode — Full dark/light theme toggle Claim Lifecycle Giver-Selects Model — Donors review all claims and choose who to give to 💬 Claim Messaging — Chat between donor and claimer (up to 5 messages per claim) 📅 Pickup Scheduling — Set a pickup date/time when approving ","listed_at":1789166861,"md_updated_at":1789166861},{"full_name":"nomeata/nanobruijn","url":"/r/nomeata/nanobruijn","github":"https://github.com/nomeata/nanobruijn","title":"nanobruijn","tagline":"Experimental Lean 4 type checker using pure de Bruijn indices","source":"trawl","tier":"found","license":"Apache-2.0","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: nanobruijn, an experimental Lean 4 type checker forked from nanodalib; \"All code modifications from the original nanodalib were written by Claude\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"nanobruijn nanobruijn -- an experimental Lean 4 type checker exploring pure de Bruijn index representation with shift-homomorphic caching. Provenance Forked from nanoda_lib by ammkrn (at commit 68d5ca9), an independent Rust implementation of the Lean 4 kernel. The original uses locally-nameless binding representation (free variables as names, bound variables as de Bruijn indices), matching the official Lean kernel. Upstream nanoda commits reviewed and ported through 6d2f037 (2026-04-07). nanobruijn replaces the binding representation and caching strategy while keeping the rest of the type-checking logic largely intact. Purpose This is a research vehicle for evaluating specific design decisions in Lean 4 kernel implementation, benchmarked against the original nanoda on real-world proof artifacts (Mathlib). It is part of the lean-kernel-arena project. This is not intended for production use. Design decisions under evaluation Pure de Bruijn indices (replacing locally-nameless) The standard Lean kernel and nanoda use locally-nameless representation: bound variables are de Bruijn indices, but when entering a binder (e.g. in whnf(forall) or infer(lambda)), the bound variable is replaced by a fresh free variable (a \"local\"). This requires a full substitution on binder entry. nanobruijn instead keeps everything as de Bruijn indices throughout. Entering a binder is free (no substitution), but looking up a bound variable in the environment requires adjusting (shifting) the retrieved va","listed_at":1789166859,"md_updated_at":1789166859},{"full_name":"iliawerner/Fixer-app","url":"/r/iliawerner/Fixer-app","github":"https://github.com/iliawerner/Fixer-app","title":"Fixer-app","tagline":"Free, open-source macOS text actions and Gemini-powered dictation via global shortcuts.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["productivity"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Fixer, a macOS menu-bar app for shortcut-driven text rewriting and dictation; the owner says 'This project was largely \"vibecoded\" with AI tools'. It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Fixer Current beta: 0.4.1 · Changelog · Roadmap Fixer is a tiny macOS menu-bar app that rewrites selected text and turns speech into text with Google Gemini. Select text in any app, press a global keyboard shortcut, and the selection is replaced with an LLM-polished version — fix grammar, translate, summarize, or whatever your prompt says. Or trigger the permanent Dictation Action and speak directly into the field that already has your cursor. No browser round-trip. A typical use: \"fix the grammar and make it sound natural.\" Fixer turns that request into a single keystroke, anywhere. Design: Fixer offers warm paper and warm charcoal themes, with signal yellow accents and quiet dividers. During a run, a compact neutral status card uses familiar progress, success, and error symbols with direct copy. It never repeats the Fixer wordmark or takes focus from the active app. A layered version of the yellow Fixer poster introduces Fixer 0.2 on first launch and can be replayed from the menu-bar menu. The Actions workspace opens as a compact 820 × 720 native window with a 40 pt sidebar titlebar and a deliberately taller 100 pt Action masthead, plus compact non-shifting pointer feedback. Design work and visual QA must use the curated design-reference package. It records the approved hierarchy, palette, identity master, intentional native adaptations, and concepts from the original prototype that must not be copied into the product. ✨ Features Safe in-place rewrite — select text in an ac","listed_at":1789166857,"md_updated_at":1789166857},{"full_name":"bytecube/DeDox","url":"/r/bytecube/DeDox","github":"https://github.com/bytecube/DeDox","title":"DeDox","tagline":"A self-hosted, privacy-first document ingestion service that captures, processes, extracts metadata, and archives documents to Paperless-ngx.","source":"trawl","tier":"found","license":"Apache-2.0","stars":5,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: DeDox, a self-hosted document capture and archiving service for Paperless-ngx; 'This project was developed using AI-assisted \"vibe coding\" methodology'. It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"DeDox - Detoxify Your Document Chaos Alpha Release Notice: This project was developed using AI-assisted \"vibe coding\" methodology with Claude (Opus 4.5) by Anthropic. While functional and tested, users should evaluate suitability for their use case. Use at your own risk. DeDox = de (undo, remove) + dox (documents) — detoxify your paperwork by transforming chaotic piles of documents into organized, searchable, AI-enhanced archives. A self-hosted, privacy-first document ingestion service that captures, processes, extracts metadata, and archives documents to Paperless-ngx. See It In Action 📱 Capture & Process Scan documents with your phone using live edge detection, then watch DeDox automatically extract metadata via local LLM. 📦 Archive & Search Documents are enriched with AI-extracted metadata in Paperless-ngx, then synced to Open WebUI for RAG-powered chat. 100% Local - Zero Internet Required Your documents never leave your network. DeDox is designed from the ground up to operate entirely within your local network: No cloud dependencies: All processing happens on your hardware No external API calls: LLM inference runs locally via Ollama or llama.cpp No telemetry or tracking: Your data stays yours Air-gapped compatible: Works without any internet connection Mobile app: Connects only via local network (LAN/WiFi) — no internet traffic, no external servers This makes DeDox ideal for sensitive documents like financial records, medical documents, legal paperwork, and pers","listed_at":1789166856,"md_updated_at":1789166856},{"full_name":"AdrianVollmer/FluxFeed","url":"/r/AdrianVollmer/FluxFeed","github":"https://github.com/AdrianVollmer/FluxFeed","title":"FluxFeed","tagline":"A modern, resource-efficient RSS feed reader built with Rust.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: FluxFeed, a self-hosted RSS reader in Rust with reader mode and search, whose README says \"This was entirely vibe-coded as part of some personal experimentation\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"FluxFeed A modern, resource-efficient RSS feed reader built with Rust. ImportantDisclaimer: This was entirely vibe-coded as part of some personal experimentation with this new technology. I know neither Rust nor more than the most basic part of CSS. If that is a deal breaker for you, I understand. Features Progressive Web App Assign groups and tags to feeds Reader mode Search The primary use case is running this program as a self-hosted Docker instance accessed locally or via VPN. There is currently no support for multiple users. Demo Try it out with Docker (or Podman): docker run --rm --init -p 3000:3000 \\ ghcr.io/adrianvollmer/fluxfeed:latest Tech Stack Web Framework: Axum 0.7 Templates: Askama (compile-time checked) Interactivity: HTMX Database: SQLite with SQLx RSS Parser: feed-rs UI: Tailwind CSS Background Jobs: tokio-cron-scheduler Development Prerequisites Rust (1.92+) Node.js (for Tailwind CSS) pkg-config and libssl-dev (for OpenSSL) Building Use the unified build script: # Build everything (frontend + Rust) ./scripts/build.sh # Or build specific targets: ./scripts/build.sh frontend # CSS, JS, and icons ./scripts/build.sh css # Tailwind CSS only ./scripts/build.sh js # TypeScript only ./scripts/build.sh icons # PNG icons from SVG ./scripts/build.sh rust # Rust debug build ./scripts/build.sh rust-release # Rust release build ./scripts/build.sh clean # Remove build artifacts Or manually: npm install npm run build # Build CSS and JS cargo build --release Run the applica","listed_at":1789166854,"md_updated_at":1789166854},{"full_name":"Mik-pe/Katla","url":"/r/Mik-pe/Katla","github":"https://github.com/Mik-pe/Katla","title":"Katla","tagline":"A pet vulkan renderer","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":612,"vulnerable":10},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Katla, a pet Vulkan game engine in Rust; asked \"Is this vibecoded?\" its README answers \"It sure is, I ain't got time to write all of this\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Katla ✨🎮 A Vulkan game engine in Rust. Playground for graphics experiments. 🐒 What's Inside 📦 Vulkan 1.3 🔺 - Dynamic rendering, Synchronization2, VMA integration, bindless textures Custom ECS 🧩 - Sparse set storage, query system, component derive macros Render graph 📊 - Resource lifetime management, automatic barrier insertion PBR materials 💎 - Hot reload support, template-based definitions WGSL shaders ✨ - Compiled via naga at runtime GLTF support 🦊 - Skeletal animation, PBR materials, background loading Editor UI 🖼️ - Declarative dockable panels, asset browser, entity inspector, transform gizmos, CodeEditor with syntect highlighting Bindless textures 🎨 - Single texture array for UI rendering, texture switching via vertex indices, no push descriptor overhead Text pipeline 🔤 - cosmic-text with HarfBuzz shaping, BiDi, CJK, word wrapping, font fallback; swash rasterization, etagere atlas packing, subpixel positioning GPU-instanced UI ⚡ - Instanced rendering (shared unit quad + per-instance data) replaces per-quad vertex emission; incremental Taffy layout caching via dirty flags Crates 📚 Crate Description katla_gfx Vulkan wrapper, render graph, materials katla_ecs Entity component system katla_math SIMD math library katla_ui Declarative UI system — Widget trait, focus chains, dockable panels, cosmic-text pipeline, GPU-instanced rendering, CodeEditor (622 tests) katla_app Application framework, components, systems Running 🏃 cargo run # Run the demo 🎮 cargo run -- -s","listed_at":1789166853,"md_updated_at":1789166853},{"full_name":"rubidus-api/hgfloater","url":"/r/rubidus-api/hgfloater","github":"https://github.com/rubidus-api/hgfloater","title":"hgfloater","tagline":"HGFloater is a lightweight, high-performance quick launcher and task switcher for Windows 11 and above. Built exclusively with pure C and WinAPI, it offers a fast and responsive experience through translucent floating boxes and optimized keyboard/mouse interactions.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tiny pure-C Win32 launcher and task switcher for Windows 11; the author lists the method as 'developed with AI assistance, in the \"vibe coding\" style'. It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"HGFloater English | 한국어 v0.17.13 — built 2026-09-11 Download hgfloater.exe — v0.17.13, 650 KB · All releases HGFloater is a lightweight desktop utility for Windows 11 and above. A small translucent widget floats on your desktop; clicking it opens a dashboard that launches your shortcuts, switches between running windows, and puts volume, brightness, opacity, and a command console one click away. It is written in pure C against the Win32 API with zero external dependencies: the whole program is a single executable of about 650 KB that needs no installer and no runtime, so it starts instantly and stays out of your way. Table of Contents Overview Install and First Run The Two Windows The Floater The Taskbox The Toolbar The Options Menu The Command Box Monitor Thumbnails Notes Keyboard Reference Mouse Reference Configuration File Files and Directories Building From Source Project Layout Tests and Verification About the Developer The HellGates Series License 1. Overview HGFloater aims at one thing: making everyday desktop control faster than the stock taskbar allows. It works as a quick launcher, a task switcher, a system control panel (volume, per-display brightness and scaling, window opacity, screen lock), a command box for driving windows by name and number, and a note pad - all reachable from a widget you can park anywhere on any monitor. Design principles worth knowing before you use it: Nothing runs in the background but the app itself. No services, no installer. A single h","listed_at":1789166849,"md_updated_at":1789166849},{"full_name":"anupshinde/goeph","url":"/r/anupshinde/goeph","github":"https://github.com/anupshinde/goeph","title":"goeph","tagline":"Fast Go ephemeris toolkit (BSP/SPK) validated against Python Skyfield with golden tests and published error bounds.","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a fast Go library for computing planetary positions from JPL ephemeris files; the owner says \"I had Claude build one inspired by Skyfield's approach\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"goeph A fast Go library for computing planetary positions from JPL ephemeris files. Origin & attribution goeph is a Go library inspired by Skyfield, the excellent Python library by Brandon Rhodes for computing positions of stars, planets, and satellites. Skyfield is the reference implementation. All goeph outputs are validated against Skyfield. goeph is not affiliated with, endorsed by, or a replacement for the Skyfield project. The mathematical foundations (IAU nutation/precession models, Chebyshev polynomial evaluation, SPK binary format) are from published standards (IERS Conventions, JPL/NAIF documentation). If you need authoritative astronomy or Skyfield's full feature set, use Skyfield directly. Why this exists I needed fast planetary position computation in Go for a research project. Skyfield is great but Python was too slow for my batch workloads. I couldn't find a Go library that did what I needed, so I had Claude build one inspired by Skyfield's approach. It's ~14x faster than the Python equivalent for the same computations (with the default 30-term nutation; ~2-3x faster with full 1365-term nutation matching Skyfield's precision — see Nutation precision). I'm publishing this because someone else is probably looking for the same thing I was. This project was coded by AI and validated against Skyfield using golden tests and end-to-end CSV comparison. What it does Loads JPL DE-series ephemeris files (tested with DE440s; supports SPK Type 2 and Type 3 segments) Com","listed_at":1789166847,"md_updated_at":1789166847},{"full_name":"dhoconno/lungfish-genome-explorer","url":"/r/dhoconno/lungfish-genome-explorer","github":"https://github.com/dhoconno/lungfish-genome-explorer","title":"lungfish-genome-explorer","tagline":"An opinionated app for working with genomic data.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a macOS app for analyzing and visualizing genomic sequencing data whose README heading reads: Lungfish Genome Explorer is unabashedly \"vibe coded\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Lungfish Genome Explorer Project site: https://dhoconno.github.io/lungfish-genome-explorer/ · User manual: https://lungfish-genome-explorer.readthedocs.io/en/latest/ · Download: Releases TL;DR I originally conceived of Lungfish Genome Explorer (LGE) as my love letter to Mac-based graphical sequence analysis software. Now I think it is more of a college-era mix tape: blending together a bunch of disparate things I like into what I hope is a thoughtful and opinionated whole. There are two versions of LGE, a more stable version that I update every few weeks and preview versions that change more frequently. LGE only runs on Apple Silicon Macs. LGE makes it easy for users to interact with many state-of-the-art open source bioinformatics tools that most users have to run from the command line. Most of the functionality is geared towards analysis and visualization of genomic data (Illumina, Pacific Biosciences, Oxford Nanopore) collected by labs and in NCBI SRA. LGE projects are designed for maximal reproducibility, can be shared easily with others, and can be extended by using agentic LLMs to perform analyses that are not built-in. For example, if you have a bunch of FASTQ files and you want to map them to a reference using a tool that isn't included, you can provide the path of the LGE project to Claude Code or Codex, and then have these tools run the mapping on data within the LGE project and save the results within LGE. I also encourage others to create their own LGE forks to me","listed_at":1789166845,"md_updated_at":1789166845},{"full_name":"maximdr86/tikpilot","url":"/r/maximdr86/tikpilot","github":"https://github.com/maximdr86/tikpilot","title":"tikpilot","tagline":"Self-hosted web panel for a MikroTik fleet: monitoring, bulk RouterOS operations, backups, syslog. Python, no build step","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted web panel for managing a fleet of MikroTik routers whose README says \"The code was written by Claude, an AI model by Anthropic.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Tikpilot English · Русский Central management for a fleet of MikroTik routers over the RouterOS API: monitoring, bulk upgrades, scheduled backups, syslog, WireGuard. Self-hosted, no cloud, MIT. Runs on any server with Python, as a single process. No Redis, no Celery, no Node.js, no frontend build step. The panel itself needs no internet access: nothing is loaded from third-party sites. The interface comes in English and Russian, switchable from the menu. The code was written by Claude, an AI model by Anthropic. See How this project came about. The prompt and the project context behind it: docs/prompt-mikrotik-panel.en.md. Site: tikpilot.ru Devices. Filters, search, sorting, RouterOS version, uptime and mobile operator for each site Monitoring. Fleet map, uptime over 24 h, latency and loss Scripts. A library of commands and what actually sits on the routers Device log. Syslog from the whole fleet, with filters and hiding rules One action for the whole fleet. Reboot, upgrade, a script, a backup, an API call Jobs. History with per-device success and failure counts Device card. Status, clients behind the router, latency and loss WireGuard. Router-to-router links through a hub, with handshakes and traffic More screenshots: clients, backups, groups, audit log, rights, ports and services The screenshots are taken from a live fleet with screenshot mode on: site names, addresses, client names and accounts are replaced with made-up ones by the panel itself. Addresses come from the rang","listed_at":1789166843,"md_updated_at":1789166843},{"full_name":"suryakosaraju/hotgaze","url":"/r/suryakosaraju/hotgaze","github":"https://github.com/suryakosaraju/hotgaze","title":"hotgaze","tagline":"Attention heatmaps for UI screenshots, with numeric scores and A/B compare, not just pretty pictures.","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tool that makes attention heatmaps for UI screenshots, \"Built with Claude Code & Codex as orchestrators and DeepSeek as the implementing agent.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"HotGaze ⏿ Attention heatmaps for UI screenshots — with numeric scores and A/B compare, not just pretty pictures. Predict where users' eyes land on a design, get a machine-readable attention share for any region, and diff two variants to see which one wins and by how much. Local, offline, MIT-licensed. pip install hotgaze && hotgaze run screenshot.png Status: v0.1 alpha — on PyPI. The fast heuristic backend, region scoring, A/B compare, and deep UNISAL backend are shipping. Faces layer (YuNet) is available via --layers faces. API and CLI may change before v1. Why this exists Every existing attention-prediction tool — paid (HeatScope, Attention Insight, EyeQuant) and free (Foveacast) — stops at a colored overlay for a human to eyeball. That's fine for a designer squinting at a mockup. It's useless when you want to: Compare two design variants and quantify which one draws more attention to the CTA. Wire attention checks into CI (\"your button just lost 23% of its attention\"). Script or automate any part of the design-review loop. HotGaze outputs the picture too, but the picture isn't the point. The numbers are. Install pip install hotgaze # core (fast backend) pip install \"hotgaze[deep]\" # + deep UNISAL backend Python 3.10+. Runs on macOS and Linux. No cloud, no API keys, no telemetry. From source git clone https://github.com/suryakosaraju/hotgaze cd hotgaze pip install -e . Quickstart # Generate an attention overlay hotgaze run screenshot.png -o overlay.png # Score a spe","listed_at":1789166842,"md_updated_at":1789166842},{"full_name":"notclickable-jordan/podcastdn","url":"/r/notclickable-jordan/podcastdn","github":"https://github.com/notclickable-jordan/podcastdn","title":"podcastdn","tagline":"A self-hosted YouTube video/playlist to podcast generator that uploads to S3 and CloudFront","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted YouTube-to-podcast feed generator whose README says \"This project was largely written by Claude Opus 4.6.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"PodCastDN Podcast -> CDN = PodCastDN A self-hosted YouTube video/playlist to podcast generator that uploads to S3 and CloudFront. WARNING: This project was largely written by Claude Opus 4.6. Be careful trusting it with your credentials. Features Multi-user authentication — Email/password with OAuth support Multiple podcast feeds — Create and manage independent feeds YouTube integration — Add individual videos or entire playlists RSS 2.0 compatible feeds S3 + CloudFront storage with custom domain support Dark mode with system preference detection Docker deployment ready Getting started Set up AWS and get the access and secret keys Configure environment variables Run the server Screenshots Version history 1.0.1 Improved mobile view","listed_at":1789166840,"md_updated_at":1789166840},{"full_name":"vinnie291/sunnylink-wiki","url":"/r/vinnie291/sunnylink-wiki","github":"https://github.com/vinnie291/sunnylink-wiki","title":"sunnylink-wiki","tagline":"A comprehensive, community-driven database and wiki for Sunnypilot settings and driving models","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a mobile-first wiki for sunnypilot settings and driving models whose README says \"This whole thing was vibecoded with Antigravity using Gemini 3 Pro & Opus 4.5\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"☀️ Sunnylink Wiki The \"Missing Manual\" for Sunnypilot. A community-driven database decoding the complex settings, toggles, and driving models of the sunnypilot ecosystem. Check it out here: https://sunnylink.wiki 🚀 Overview Sunnypilot (a fork of openpilot) offers incredible customization, with over 100+ toggles and 60+ driving models. However, for new users, understanding the difference between \"MADS\" and \"NNLC\" or choosing between \"WMI v12\" and \"Recertified Herbalist\" can be overwhelming. Sunnylink Database solves this by acting as the single source of truth. It is a mobile-first web app that translates technical settings into plain English, helping drivers configure their Comma devices with confidence. ✨ Key Features 🎛️ Toggle Encyclopedia: A searchable wiki for every setting (MADS, Dynamic Experimental Control, etc.) with definitions, safety warnings, and dependencies. 🧠 Model \"Vibe Check\" Explorer: A breakdown of driving models not just by version, but by feel (e.g., \"Aggressive,\" \"Comfort,\" \"Legacy\"). 📱 Mobile-First Design: Optimized for in-car reference, allowing you to look up settings while sitting in your driver's seat. 🔌 Antigravity Ready: All data is stored in structured JSON, serving as the backend for the Sunnylink \"Antigravity\" app ecosystem. 🧠 The Models Page (Model Explorer) Navigating the dozens of available driving models can be confusing. Names like \"WMI v12\" or \"Dark Souls\" don't explain how the car will actually behave. The Models Page solves this b","listed_at":1789166838,"md_updated_at":1789166838},{"full_name":"Neoxra/vscode-bt-viewer","url":"/r/Neoxra/vscode-bt-viewer","github":"https://github.com/Neoxra/vscode-bt-viewer","title":"vscode-bt-viewer","tagline":"Behavior Tree cpp viewer ","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a VS Code extension for viewing and live-monitoring BehaviorTree.CPP trees whose README says \"This extension was built with Claude Code (Anthropic).\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"BehaviorTree Viewer Early access -- actively in testing, expect bugs. Feedback welcome. A free, open-source VSCode extension for visualizing and monitoring BehaviorTree.CPP v4 XML files. A lightweight alternative to Groot2. Horizontal (small trees) Waterfall (large trees) Features Tree Visualization Parses BehaviorTree.CPP v4 XML and renders an interactive tree diagram Color-coded nodes by category: Control (orange), Decorator (green), Action (blue), Condition (yellow), SubTree (purple), Script (grey) Drag nodes to rearrange the layout; overlapping nodes auto-resolve Collapse/expand subtrees with double-click Inline SubTree expansion Blackboard Inspector Lists all {variable} references found in the tree Shows which nodes read and write each variable Node Palette Displays all node types from the TreeNodesModel section Grouped by category with port definitions (name, type, direction, defaults) Search Filter nodes by name, type, or port values Matching nodes highlighted; non-matches dimmed Live Monitor (requires ZMQ) Connects to a running BT.CPP process via ZMQ (same protocol as Groot2) Real-time node status overlay: Running (pulsing blue), Success (green), Failure (red), Idle (dimmed) Loads the live tree with correct UIDs for accurate status mapping Multi-tree Support Dropdown selector for XML files containing multiple <BehaviorTree> definitions Usage Open any BehaviorTree.CPP v4 XML file in VSCode Open the viewer using any of: Editor title bar: click the tree icon (top-r","listed_at":1789166837,"md_updated_at":1789166837},{"full_name":"liuyxpp/AcceleratedDCTs.jl","url":"/r/liuyxpp/AcceleratedDCTs.jl","github":"https://github.com/liuyxpp/AcceleratedDCTs.jl","title":"AcceleratedDCTs.jl","tagline":"Device agnostic implementation of 1D/2D/3D DCTs.","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"Julia","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a device-agnostic DCT library for Julia on CPU and GPU whose README says \"Most of source codes and docs in this project are generated by Claude Opus 4.5\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"AcceleratedDCTs.jl Fast, Device-Agnostic, AbstractFFTs-compatible DCT library for Julia. AcceleratedDCTs.jl provides highly optimized Discrete Cosine Transform implementations for 1D, 2D and 3D data: DCT-II (Standard \"DCT\") and DCT-III (Inverse DCT) DCT-I and IDCT-I (symmetric boundary conditions) It leverages KernelAbstractions.jl to run efficiently on both CPUs (multithreaded) and GPUs (CUDA, AMD, etc.), and implements the AbstractFFTs.jl interface for easy integration. Key Features ⚡ High Performance: optimized algorithms (Makhoul's method) that outperform standard separable approaches. 🚀 Device Agnostic: Runs on CPU (Threads) and GPU (CuArray, ROCArray via KernelAbstractions). 🔥 VkDCT Backend: Pre-compiled VkFFT-based CUDA library (VkDCT_jll) for DCT-I offering ~15x speedup on GPU. Zero-setup: just using CUDA, AcceleratedDCTs. 🧩 AbstractFFTs Compatible: Zero-allocation mul!, ldiv!, and precomputed Plan support. 📦 3D Optimized: Specialized 3D kernels that avoid redundant transposes. Installation using Pkg Pkg.add(\"AcceleratedDCTs\") Quick Start Basic Usage using AcceleratedDCTs: plan_dct, mul! using CUDA # 1. Create Data N = 128 x_gpu = CUDA.rand(Float64, N, N, N) # can be any Real, e.g. Float32 # 2. Create Optimized Plan (Recommended) p = plan_dct(x_gpu) # 3. Execute y = p * x_gpu # Standard execution mul!(y, p, x_gpu) # Zero-allocation (in-place output) # 4. Inverse x_rec = p \\ y # or inv_p = inv(p) mul!(x_rec, inv_p, y) One-shot Functions For convenience (slower due ","listed_at":1789166836,"md_updated_at":1789166836},{"full_name":"torresjeff/rtmp-rs","url":"/r/torresjeff/rtmp-rs","github":"https://github.com/torresjeff/rtmp-rs","title":"rtmp-rs","tagline":"A modern RTMP server in Rust with HEVC/AV1 support that handles encoder quirks, late joiners, and slow subscribers so you don't have to.","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["media"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an async RTMP streaming server and client library in Rust whose README says \"Almost all of the code in this Rust version was written by AI (Claude Opus 4.5).\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"rtmp-rs: RTMP server and client library for Rust rtmp-rs is an async RTMP server and RTMP client for Rust, built on Tokio. Use it to ingest live streams from OBS or FFmpeg, relay them to viewers, pull a stream from another server, or publish to YouTube, Twitch and other RTMPS endpoints. It supports legacy RTMP (H.264 and AAC) as well as Enhanced RTMP v2 with HEVC, AV1, VP9, Opus and FLAC. The server accepts what real encoders send. Empty app names, timestamp regressions and other quirks are tolerated instead of dropping the connection. A GOP cache gives late joiners a keyframe right away, and slow subscribers lose video frames before they lose audio. What you get An RTMP server with stream key routing and pub/sub built in. Publishers and players are matched by stream key with no extra code. An RTMP client that can pull a stream or publish one, over plain RTMP or RTMPS (TLS). Enhanced RTMP v2 with codec negotiation. HEVC, AV1, VP9 and VP8 video, Opus, FLAC, AC-3 and E-AC-3 audio, plus H.264 and AAC. Older clients fall back to legacy RTMP automatically. A RtmpHandler trait with optional callbacks for authentication, metadata and per-frame access. Every callback has a default, so you override only the ones you care about. GOP caching so a viewer joining mid-stream starts on a keyframe instead of waiting for the next one. Backpressure handling that drops video for slow subscribers while keeping audio continuous. Media payloads passed around as bytes::Bytes, so a frame is shared b","listed_at":1789166834,"md_updated_at":1789166834},{"full_name":"gtfierro/shifty","url":"/r/gtfierro/shifty","github":"https://github.com/gtfierro/shifty","title":"shifty","tagline":"Fast vibe-coded shacl validation+inference","source":"trawl","tier":"found","license":"BSD-3-Clause","stars":8,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a SHACL validation and inference engine in Rust with Python and WebAssembly builds, described as \"Fast vibe-coded shacl validation+inference\". It carries the BSD-3-Clause license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"shifty A formalism-first SHACL validation and SHACL-AF inference engine written in Rust, grounded in the algebraic treatment of Common Foundations for SHACL, ShEx, and PG-Schema (arXiv:2502.01295). Available as a command-line tool, Python bindings (pyshifty), a C++17 static-library SDK, and a WebAssembly module that runs in the browser. See the archive-first-attempt branch for an older attempt, which was published as shifty 0.0.7 Disclosure: This project was nearly 100% \"vibe coded\" with Claude Opus 4.8, Sonnet 4.6, and ChatGPT 5.5 Features Full SHACL Core validation — node and property shapes, all standard constraint components SHACL-AF inference — forward-chaining sh:rule evaluation (Triple Rules, SPARQL Construct Rules) to a fixed point, with stratification analysis for recursive rulesets Algebraic IR — shapes are lowered to a path algebra (π) and shape grammar (φ) before evaluation; the same IR drives both validation and inference Native SPARQL execution — a subset of sh:sparql constraints and SPARQL Construct rules runs directly over an indexed dataset without a full SPARQL engine, with automatic fallback to Spareval for unsupported constructs Multi-layer pipeline — parsing → algebraic lowering → normalization/CSE → physical planning → execution; each layer is independently inspectable pyshifty-compatible Python API — validate() returns (conforms, report_graph, results_text) matching pyshifty's interface Feature support Legend: ✅ supported · ⚠️ partial / gated · ❌ unsupp","listed_at":1789166833,"md_updated_at":1789166833},{"full_name":"mdemirhan/dux","url":"/r/mdemirhan/dux","github":"https://github.com/mdemirhan/dux","title":"dux","tagline":"A disk usage tool written in Python","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a parallel disk usage analyzer with an interactive TUI whose README says \"100% AI-written.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"dux — a fast, pretty, interactive disk usage analyzer Parallel disk usage analyzer for macOS and Linux. Scans directories with multi-threaded I/O, categorizes files (temp, cache, build artifacts), and presents results as rich CLI tables or an interactive TUI with vim-style navigation. 100% AI-written. The vast majority of this codebase was written by Claude (Anthropic), with contributions from Codex (OpenAI). Human involvement was limited to directing, reviewing, and benchmarking. Screenshots CLI summary (uv run dux -n) TUI overview (uv run dux) TUI browse — expandable directory tree with disk usage bars Features Parallel scanning with configurable thread pool (default 4 workers) Interactive TUI with 5 views, vim keybindings, localized subtree refresh, search/filter, and pagination Composable CLI flags — --top-temp, --top-cache, --top-dirs, --top-files each print their own table and can be freely combined 59 built-in pattern rules for detecting temp files, caches, and build artifacts across dozens of ecosystems (Node, Python, Rust, Go, JVM, Swift, C++, and more) Fully configurable via JSON config with pattern overrides and custom paths Analysis-only — never deletes, moves, or modifies your files Quick Start Requires Python 3.13+ and uv. # Clone and install git clone https://github.com/mdemirhan/dux.git cd dux uv sync # Analyze current directory in the interactive TUI uv run dux # Analyze a specific path interactively uv run dux ~/src # Print a non-interactive summary and exit","listed_at":1789166831,"md_updated_at":1789166831},{"full_name":"mehmetefeumit/Haven-App","url":"/r/mehmetefeumit/Haven-App","github":"https://github.com/mehmetefeumit/Haven-App","title":"Haven-App","tagline":"Haven is a private and censorship-resistant location sharing app for Android and iOS, running on Nostr using the Marmot protocol.","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"Dart","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an end-to-end encrypted location sharing app built on Nostr whose README says \"Haven was written entirely using AI.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"IMPORTANT: Haven is still in active development. The following still needs to be done before it is ready for a 1.0 release: 3rd party security audit (see AI Disclaimer below.) Beta tests and incorporating user feedback. Allow better usage of existing Nostr keys (i.e. list of invitees from the following list), and signer support. Nostr-agnostic UI/UX improvements. ... Haven Private and censorship-resistant location sharing. Haven is an end-to-end encrypted location sharing app that requires no personal information at sign-up and uses no central server. Your location is encrypted on your device before it is sent to decentralized Nostr relays. It can only be decrypted by members of your location sharing circle after they fetch your latest location information from the relays. You can select the relays you want to send location updates through, and/or run and use your own relays. Haven requires no phone numbers, emails, or any personally identifiable information before you can start using it. It is open-source, so all of its privacy guarantees are verifiable. Why use Haven? Most technology companies we trust our most personal information with have not proven their interest in keeping our data secure and private. Instead, our data is commodified: even entities which we do not directly interact with buy and sell it to build a profile of who we are, where we go, and what we care about the most. For example, one of the most popular family-location apps was found selling the precise l","listed_at":1789166830,"md_updated_at":1789166830},{"full_name":"RhetTbull/gittwig","url":"/r/RhetTbull/gittwig","github":"https://github.com/RhetTbull/gittwig","title":"gittwig","tagline":"TUI CLI app for viewing and switching git branches","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a terminal UI for browsing and switching git branches whose README says \"Built with Claude Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"gittwig A terminal user interface (TUI) for viewing and switching git branches, built with Python and Textual. I built this because I wanted a quick way to view git branches and switch between them. Built with Claude Code. This provides a fast, easy way to browse branches, see the commit history for a given branch, identify files changed in a branch, and view file diffs with syntax highlighting. As an open source developer who works with many branches across multiple projects, this is very helpful for figuring out where I left off and juggling multiple branches. Features Browse local and remote branches in a navigable list View changed files compared to your default branch View commit history unique to each branch Syntax-highlighted diff viewer Checkout branches with Enter Create and delete branches Fetch, push, and pull operations Filter branches with / search Vim-style keyboard navigation Installation Requires Python 3.12 or later. Install uv and run: uv tool install gittwig To run with uv without installing: uv tool run --from gittwig twig Usage Run twig in any git repository: twig Or specify a repository path: twig /path/to/repo Keyboard Shortcuts Navigation Key Action j / k Move cursor down / up h / l Focus left / right pane g g Go to top of list G Go to bottom of list Ctrl+d / Ctrl+u Page down / up Enter Select item / checkout branch Branch Operations Key Action n Create new branch and switch to it d Delete branch (with confirmation) r Refresh data f Fetch from remotes ","listed_at":1789166828,"md_updated_at":1789166828},{"full_name":"MaxBerktoldRWTH/BRICKbuilder","url":"/r/MaxBerktoldRWTH/BRICKbuilder","github":"https://github.com/MaxBerktoldRWTH/BRICKbuilder","title":"BRICKbuilder","tagline":"Visual editor for the fast creation and distribution of BRICK Graphs (https://brickschema.org/)","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a visual editor for Brick building-system graphs whose README says \"The complete code, including this readme was written by claude 3.7!\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Building Systems Design Tool This application provides a graphical interface for designing building systems using the Brick and REC ontologies. It allows you to create, visualize, and export semantic models of building systems with entities, connections, and their relationships. Core Functionality Entity Modeling Entity Library: Browse and use a comprehensive collection of Brick and REC entity types organized by category (Equipment, Sensors, Points, Spaces, etc.) Drag-and-Drop Interface: Easily add entities to your design by dragging them from the entity browser to the canvas Properties Panel: View and edit entity properties, labels, and positions Semantic Relationships Visual Connections: Create semantically meaningful connections between entities with proper Brick relationship types Smart Routing: Use joints to create custom routing paths for connections Relationship Types: Automatically assigns appropriate relationship types (e.g., hasPoint/isPointOf for sensor connections) Design Management Export to Turtle: Save your designs as standard RDF Turtle (.ttl) files for use with other semantic tools Import from Turtle: Load previously saved designs back into the application IFC Import: Import building layouts directly from IFC files via drag-and-drop Keyboard Shortcuts Shortcut Function H Show help window with all shortcuts Del Delete selected items Ctrl+A Select all items R Rotate selected entities 90° Ctrl+C Copy selected items Ctrl+V Paste copied items P Toggle visibility o","listed_at":1789166827,"md_updated_at":1789166827},{"full_name":"shortusername/projectivy-weather-wallpaper","url":"/r/shortusername/projectivy-weather-wallpaper","github":"https://github.com/shortusername/projectivy-weather-wallpaper","title":"projectivy-weather-wallpaper","tagline":"Weather panel wallpaper plugin for Projectivy Launcher","source":"trawl","tier":"found","license":"Apache-2.0","stars":10,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a weather wallpaper plugin for the Projectivy Android TV launcher whose README says \"Most of the Kotlin in this repo was written by Claude\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Weather Wallpaper for Projectivy A wallpaper provider plugin that turns your Projectivy Launcher home screen into a weather display. Current conditions, optional forecast panels, and backgrounds that react to the weather. Weather data from Open-Meteo. No API key, no account, no tracking. Install Easiest — no computer needed. Install Downloader by AFTVnews on the TV, then enter code 6704396. That pulls the latest release directly and will keep pointing at future versions too. Or side-load the APK from the latest release by USB or adb. Or build it yourself — recommended, and it takes about two minutes with no local toolchain. Fork the repo, go to Actions → Build APK → Run workflow, and download the artifact. The released APK is debug-signed, so building it yourself means not having to trust a stranger's binary. Then in Projectivy: Settings → Appearance → Wallpaper → Launcher wallpaper → Plugins → Weather Wallpaper. On first run the plugin estimates your location from your IP, so it works immediately. Correct it in settings via the gear icon beside the plugin name — locations are added by name, not coordinates. What it shows Always: temperature, conditions, feels-like, today's high and low, and wind with direction. Each of these can be toggled independently: Panel What it adds Rain and snow timing \"Rain starting in about 25 min\", from 15-minute data. On by default. Helpful advice \"Frost likely tonight\", \"dry until 4 PM\", \"good drying day\". On by default. Hourly Next 6 hours with","listed_at":1789166825,"md_updated_at":1789166825},{"full_name":"tylerbcrawford/subgeneratorr","url":"/r/tylerbcrawford/subgeneratorr","github":"https://github.com/tylerbcrawford/subgeneratorr","title":"subgeneratorr","tagline":"Generate and translate subtitles for Plex, Jellyfin, and Emby — Deepgram Nova-3 or fully local Whisper (offline, $0). Keyterm prompting for proper nouns, LLM translation into 54 languages (Claude/GPT/Gemini/Ollama), Docker Web UI + CLI.","source":"trawl","tier":"found","license":"MIT","stars":10,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a subtitle generator for Plex, Jellyfin and Emby whose README says \"Built with Claude Code.\" and \"The agent wrote most of the code under my direction.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Subgeneratorr Subtitle generation for Plex, Jellyfin, and Emby — cloud or fully local. Transcribe with Deepgram Nova-3 or an offline Whisper engine on your own hardware, then let Claude, GPT, Gemini, or a local Ollama model handle translation into 54 languages. Built with Claude Code. I set the requirements and the architecture, made the design decisions, and run it in production on my media server, with real users. The agent wrote most of the code under my direction. Subgeneratorr generates language-tagged SRT subtitles for the media Bazarr can't find — obscure shows, older episodes, anything without community subs. It runs as a Docker container with a Web UI and CLI, uses keyterm prompting so proper nouns come out right, and can translate every subtitle it makes into 54 more languages with the original timing preserved. I created it to fill the hundreds of missing-subtitle gaps in my own media library. Cloud or local — you choose Since v3, transcription is a per-job switch right on the main screen: ☁️ Cloud — Deepgram Nova-3 🏠 Local — Whisper (since v3) Accuracy Best in class; keyterms boost proper nouns up to ~90% ~89% word-level agreement with Nova-3 (small model) Speed ~1% of runtime Near-real-time on a modest CPU (0.3× measured on an N100) Cost ~$0.0057/min, $200 free credits (~585 hrs) $0, forever Privacy Audio goes to Deepgram's API Nothing leaves your machine Setup API key No account, no key; larger image with the model baked in Extras Diarization, redaction, audio ","listed_at":1789166824,"md_updated_at":1789166824},{"full_name":"blakewatson/scroll-my-mac","url":"/r/blakewatson/scroll-my-mac","github":"https://github.com/blakewatson/scroll-my-mac","title":"scroll-my-mac","tagline":"Accessibility tool for scrolling in macOS","source":"trawl","tier":"found","license":"MIT","stars":11,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A macOS accessibility app for click-and-drag scrolling anywhere, made for people who cannot use a trackpad; \"This app was built 100% with Claude\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Scroll My Mac A macOS accessibility app that lets you scroll anywhere by clicking and dragging. Vibe code alert This app was built 100% with Claude, Anthropic's AI assistant, and GSD. I don't know Swift. I wouldn't have made this app without AI, but it's something I wish existed. Use at your own risk. What It Does Scroll My Mac adds system-wide click-and-drag scrolling to macOS. Toggle scroll mode with a hotkey, then click and drag to scroll any scrollable area on your screen. System-wide scroll mode. Activated via configurable hotkey (default: F6). Menu bar icon. Access settings and toggle scroll mode from the menu bar. Natural inertia. Momentum scrolling with smooth deceleration when you release the drag, like iOS or a trackpad. Click-through detection. Small movements (~8px) pass through as normal clicks, so you won't accidentally scroll when you meant to click. macOS Accessibility Keyboard aware. Clicks on the macOS on-screen keyboard always pass through instantly, even while scroll mode is active. Launch at login. Optional automatic startup. Excluded apps. Prevent scroll mode from activating in specific apps without disabling scroll mode globally. Free forever. I made this because I needed it, not to make money. Why It Exists This app was built for users who cannot use a trackpad or scroll wheel due to disability. I rely on an on-screen keyboard and mouse for input, and needed a way to scroll without a scroll wheel or trackpad gesture. It was inspired by the ScrollAnywhe","listed_at":1789166821,"md_updated_at":1789166821},{"full_name":"txjmb/jsonata-core","url":"/r/txjmb/jsonata-core","github":"https://github.com/txjmb/jsonata-core","title":"jsonata-core","tagline":"Python Jsonata implementation with Rust backend also available as standalone Rust library","source":"trawl","tier":"found","license":"MIT","stars":11,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":358,"vulnerable":2},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A Rust implementation of the JSONata query language with Python bindings and a C library; \"Much of this project was built using Claude Code with significant human oversight.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"jsonata-core (rust) + jsonatapy + jsonata C-ABI library + jsonata cli Pypi stats Crates.io stats Click the chart for the full category-by-category tables, including jsonata-python and jsonata-rs, which are left off the chart because at 68 ms–9 s they would flatten everything else into a sliver. ⚠️ Upgrading to 2.2.8 \"Conform-ata\"+ — please test before you deploy 2.2.8 is primarily a conformance release: it corrects a number of places where this library disagreed with the jsonata-js reference. These issues were found during routine quality checks and documentation improvements. The test harness was inadvertently missing some edge cases. Many of those corrections change the result of expressions that already run today, so an upgrade is not guaranteed to be drop-in. A lot of work went into this release and a significant amount of test coverage was added to improve future results with the goal of maintaining the highest level of conformance to the jsonata standard. If you encounter any issues in testing that are not caused by the conformance changes or if you believe a change is in error, please file an issue and we'll work on it as soon as possible. If you use the library or the CLI, please add a star to the project! The areas most likely to affect you: Boolean coercion of containers. A one-element array holding a falsy value — [0], [\"\"], [false], [null] — is now falsy, as the reference has it. This affects ? :, and, or, $not, filter predicates and $filter. $formatNumber rounds ","listed_at":1789166820,"md_updated_at":1789166820},{"full_name":"mrebbert/1komma5-ha","url":"/r/mrebbert/1komma5-ha","github":"https://github.com/mrebbert/1komma5-ha","title":"1komma5-ha","tagline":"Home Assistant integration (HACS) for the 1KOMMA5° / 1komma5 / 1k5 Heartbeat energy platform: dynamischer Stromtarif, PV, battery, Wärmepumpe, EV wallbox, weather forecast, full Energy Dashboard support.","source":"trawl","tier":"found","license":"MIT","stars":13,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["iot"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: An unofficial Home Assistant integration for the 1KOMMA5° Heartbeat solar, battery and heat-pump platform, \"built 100 % vibe-coded with AI assistance\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"1KOMMA5° Home Assistant Integration 1KOMMA5° (also written 1komma5, 1k5) runs the Heartbeat energy platform for solar + battery + heat pump + EV wallbox setups. This is the Home Assistant integration for it, distributed via HACS at github.com/mrebbert/1komma5-ha. Dynamic electricity prices (dynamischer Stromtarif), 30-hour price forecast, AI optimization decisions, per-device cost allocation, cheapest-charging-window scheduling, weather forecast, and cloud push notifications, all as native HA sensors, services and bus events. Works on both 1KOMMA5°-backend accounts (emp_type: \"1K5\", most newer installs) and legacy GridX-backed accounts (emp_type: \"GRIDX\"). Highlights ⚡ 60+ sensors across 7 coordinators covering power, energy, cost, dynamic prices, AI decisions, weather and system health 💶 Dynamic tariff support with 30-hour price forecast, cheapest-window sensors, and per-consumer cost allocation 🤖 AI optimization events — react to Heartbeat's grid-charge / heat-pump-recommend decisions in real time 🔔 Cloud notification bridge (v0.1.52) — drive automations from the same push notifications the 1KOMMA5° mobile app receives 🌍 Seven locales (DE, NL, FI, ES, DK, SE, AU) with per-market currency (EUR, DKK, SEK, AUD) 🏠 Energy Dashboard-ready out of the box; no template sensors, no manual wiring 📥 Available in the HACS Default Store — no custom repository setup needed Table of Contents Quick start Configuration What you get Live power & battery Dynamic electricity pricing E","listed_at":1789166817,"md_updated_at":1789166817},{"full_name":"joestump/joe-links","url":"/r/joestump/joe-links","github":"https://github.com/joestump/joe-links","title":"joe-links","tagline":"Self-hosted go-links service — short memorable slugs that redirect to long URLs. Go + HTMX + DaisyUI/Tailwind. Single binary.","source":"trawl","tier":"found","license":"MIT","stars":12,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A self-hosted go-links service for short memorable redirects; its README says \"joe-links was designed and written by Claude (Anthropic's AI assistant)\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"joe-links A self-hosted \"go links\" service. Type go/slack in your browser and get redirected to your team's Slack workspace. Type go/hr and land on the HR portal. joe-links turns short, memorable slugs into instant redirects to any URL -- no browser extension required, just a single DNS entry and a search engine shortcut. Features Short memorable slugs -- [a-z0-9][a-z0-9-]*[a-z0-9], min 2 characters, globally unique OIDC authentication -- sign in with Google, Okta, Authentik, Keycloak, or any OpenID Connect provider Co-ownership -- multiple users can manage the same link REST API with Personal Access Tokens -- automate link management from scripts and CI OpenAPI / Swagger UI -- interactive API docs at /api/docs/ Dark / light / system theme -- automatic theme switching via DaisyUI Multi-database support -- SQLite (zero config), PostgreSQL, or MySQL Single binary -- one joe-links binary with embedded templates and static assets Quick Start (Docker) cp .env.example .env # Edit .env with your OIDC provider details docker compose up -d # Visit http://localhost:8080 The app inside the container always listens on port 8080. To publish it on a different host port, set JOE_HOST_PORT in .env (e.g. JOE_HOST_PORT=9000 serves http://localhost:9000) -- do not set JOE_HTTP_ADDR for Docker deployments; compose pins it to :8080 inside the container. The container runs as a non-root user (joe, uid 65532). If you are upgrading a deployment whose data volume was created by an older root-run imag","listed_at":1789166814,"md_updated_at":1789166814},{"full_name":"Krever/ssbudget","url":"/r/Krever/ssbudget","github":"https://github.com/Krever/ssbudget","title":"ssbudget","tagline":"Super Simple budget tracker focused on current situation.","source":"trawl","tier":"found","license":"MIT","stars":11,"language":"Scala","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A self-hosted budget tracker that answers how much you can spend this month; \"This project was 100% vibe-coded using Claude Code over 8 sessions (~4 days).\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"SSBudget A personal budget tracker built to answer one question: \"How much can I spend this month?\" The Problem Traditional budget apps obsess over categories and past spending. SSBudget focuses on the future. SSBudget takes a different approach: Define your planned items — the fixed expenses (rent, bills), incomes and savings top-ups you expect this period Let category budgets derive your variable spend (groceries, fuel) from real bank transactions Update account balances and mark previously planned payments throughout the period Get your remaining free cash — your budget until the next paycheck demo.webm Additional Features Multi-currency — PLN, EUR, USD + 30 more with live exchange rates Passkey auth — Passwordless login, no user management needed Import/Export — You can take your whole database with you (god bless SQLite!) Spreadsheet-like UI — Edit in place, minimal clicks, maximum density Self-hosted — You can run it yourself anywhere. E2E test coverage — Everything important is covered with e2e selenium test. WhatsApp notification - So that you can proactively share your budget with your spouse. Built With Claude This project was 100% vibe-coded using Claude Code over 8 sessions (~4 days). The entire codebase — backend, frontend, database schema, tests, deployment config — was generated through conversation with Claude Opus/Sonnet. Check docs/sessions/ for the session logs and CLAUDE.md for the context file that guided development. Tech Stack Layer Stack Language Scala","listed_at":1789166813,"md_updated_at":1789166813},{"full_name":"VitaliAndrushkevich/pulse","url":"/r/VitaliAndrushkevich/pulse","github":"https://github.com/VitaliAndrushkevich/pulse","title":"pulse","tagline":"Self-hosted uptime monitoring with 10+ protocols, real-time WebSocket updates, notifications, and an embedded web UI - all in a single Go binary.","source":"trawl","tier":"found","license":"Apache-2.0","stars":16,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":407,"vulnerable":4},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A self-hosted uptime monitor shipped as a single Go binary with an embedded web UI, which its README says was \"Vibecoded with Kiro\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Pulse Pulse is a self-hosted uptime monitoring platform. It ships as a single binary with an embedded web UI, backed by PostgreSQL and TimescaleDB for time-series storage. Designed for reliability at 500+ monitors with bounded worker pools, real-time WebSocket updates, and an API-first architecture. Pulse is an independent take on uptime monitoring — built with its own perspective on architecture and workflows. If you have questions, feature requests, or feedback, feel free to open an issue or reach out. Contributions of any kind are very welcome. Vibecoded with Kiro — an AI-powered IDE that turns ideas into working software through structured specs, steering files, and iterative development. Also there were some steps which were implemented with Copilot. Key Features Multi-protocol monitoring — HTTP/HTTPS, HTTP/3, QUIC, TCP, UDP, WebSocket, gRPC, DNS, ICMP, SMTP Notifications — Email (SMTP) and webhook channels with trigger-based rules, deduplication, retry with exponential backoff, and reminders Single deployable container — Go binary with embedded SvelteKit frontend Real-time updates — WebSocket diff/patch messages for instant UI sync API-first — full REST API with OpenAPI 3.0.3 spec Dashboard widgets — StatusRing, HealthScore, UptimeHeatmap, ResponseSparklines, SSLWarnings, IncidentsPanel, EventsFeed Internationalization — 13 languages with RTL support (Arabic), lazy-loaded locale bundles MCP server — manage monitors and get the status via AI assistants (Kiro, Copilot, Cl","listed_at":1789166811,"md_updated_at":1789166811},{"full_name":"lkernan/ABRP_Uploader","url":"/r/lkernan/ABRP_Uploader","github":"https://github.com/lkernan/ABRP_Uploader","title":"ABRP_Uploader","tagline":"A small app for the MG4 to push data from the car's API to A Better Route Planner's API without needing an OBD Dongle or Home Assistant. ","source":"trawl","tier":"found","license":"MIT","stars":16,"language":"Java","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A small Android app for the MG4 that pushes car telemetry to A Better Route Planner without an OBD dongle; its README says \"Written by Claude Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"ABRP_Uploader A small app for the SAIC Motor MG4 (eh32) to push data from the car's API to A Better Route Planner's API without needing an OBD Dongle or Home Assistant. You'll need to run the app once to set the API key. After that, it will run at startup as long as you've enabled it from the UI. You can get an ABRP API key at: https://abetterrouteplanner.com/home/app/api-keys/telemetry Written by Claude Code. Installing APKs on the MG4 If you haven't sideloaded apps before: Open the Bluetooth Settings and select the car name Hold down on the comma until the settings appear. Select Android Keyboard Settings. Select Languages Use the search box at the top to search for \"file\" That should give you access to the files app. From there you can select a connected USB device with APKs on it and install as usual.","listed_at":1789166808,"md_updated_at":1789166808},{"full_name":"hrzlgnm/mdns-tui-browser","url":"/r/hrzlgnm/mdns-tui-browser","github":"https://github.com/hrzlgnm/mdns-tui-browser","title":"mdns-tui-browser","tagline":"Cross platform TUI mdns-browser written in rust, vibe coded with opencode free model","source":"trawl","tier":"found","license":"MIT","stars":14,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":250,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A cross-platform terminal mDNS service browser in Rust, which its owner says was \"vibe coded with opencode free model\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"mDNS TUI Browser A terminal-based mDNS service browser built with Rust, using ratatui for the TUI interface. For an alternative running as a desktop app, check out mDNS-Browser. Table of Contents Features Quick Start Build Release Version Install from AUR Install using Homebrew Install using winget Install using snap CLI Options JSON State Dump Controls Architecture Project Structure Building Build Provenance Auditable Builds Immutable releases Future Enhancements License Features 🖥️ Interactive TUI: Full terminal UI with keyboard navigation 📊 Service Details: Display IP address, port, and TXT records 🔄 Real-time Updates: Async architecture for live service discovery 📈 Live Metrics: Comprehensive ServiceDaemon and application metrics 🎛️ Service Type Filtering: Browse different service types separately 🔍 Quick Filter: Text-based search across all service fields 📋 Advanced Sorting: Sort by Host, Type, Name, Port, Address, or Time in both directions ⚙️ Custom Service Discovery: Specify service types via CLI for hard-to-discover services 💾 JSON State Dump: Export complete application state to JSON with Ctrl+J Quick Start cargo run Build Release Version cargo build --release Install from AUR (Arch Linux) You can install mdns-tui-browser from the Arch User Repository (AUR) using yay: # Install using yay yay -S mdns-tui-browser This will install the latest version from the AUR and handle updates automatically with your regular system updates. Install using Homebrew (macOS) Y","listed_at":1789166807,"md_updated_at":1789166807},{"full_name":"yaroslav/audition","url":"/r/yaroslav/audition","github":"https://github.com/yaroslav/audition","title":"audition","tagline":"Auditions Ruby scripts, gems, Rack apps, and Rails applications for the ability to run under Ractors: static analysis of Ractor-isolation violations plus dynamic in-Ractor probing, with explanations and fixes.","source":"trawl","tier":"found","license":"MIT","stars":26,"language":"Ruby","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A Ruby tool that audits scripts, gems and Rails apps for Ractor compatibility; its README warns \"The entire codebase was written by Claude Fable 5 (Anthropic).\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Audition Point it at a Ruby script, a gem, a Rack app, or a Rails root and it tells you whether that code can run inside Ractors, why it cannot, and how to fix it. Unlike a linter, audition does not stop at pattern-matching your source: whole-program analysis is powered by rubydex, Shopify's Ruby indexer, and the target is also loaded in a sandboxed subprocess to observe real Ractor::IsolationErrors on the live object graph. Some of the checks and fixes were trained on how Rails core itself is being ractorized; see the pattern study. Three probes, one verdict. Per-file Prism AST checks, whole-program semantic analysis powered by rubydex (Shopify's code graph; class-level state is resolved across files and reopenings), and dynamic in-Ractor execution of the actual target. Explains, not just flags. Every finding carries a why (which rule of the Ractor model it violates) and a fix (what to write instead). --fix like RuboCop, in two tiers. Safe corrections: .freeze on string constants, sentinels, and containers whose elements are all shareable, Ractor.make_shareable(...) for the remaining mutable and shallow-frozen containers and for Proc constants, and boot-time hoisting of method-body requires. --fix-unsafe adds semantics-affecting rewrites: magic-comment insertion, freeze-on-memoize for class-level memoization (both @x ||= and return @x if defined?(@x) idioms keep their caching, the memoized value becomes shareable, Rails-core style; Ractor.store_if_absent only for block initi","listed_at":1789166805,"md_updated_at":1789166805},{"full_name":"varunpan/quicklogger","url":"/r/varunpan/quicklogger","github":"https://github.com/varunpan/quicklogger","title":"quicklogger","tagline":"Mobile-first PWA for logging fuel fillups to self-hosted LubeLogger","source":"trawl","tier":"found","license":"MIT","stars":26,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":412,"vulnerable":2},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A mobile-first PWA for logging fuel fill-ups to a self-hosted LubeLogger, \"developed with Claude Code\" where, per the owner, \"the assistant did the typing\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"quicklogger Mobile-first PWA for logging fuel fill-ups to a self-hosted LubeLogger instance. Status: v0.3.3 — stable. Single-user homelab tool, daily-driven. Public repo so anyone can fork and self-host. Built with Claude Code This project was developed with Claude Code — design, implementation, tests, infra, and docs were paired with Anthropic's coding assistant. Code is reviewed and decisions are made by a human; the assistant did the typing. Why LubeLogger's web UI is great for review and analytics, but entering a fill-up at the gas pump from a phone is fiddly. quicklogger is a one-form, install-as-PWA front door optimised for the pump: Auto-selects the last vehicle Volume in gallons or liters, cost in any major currency — converted server-side Live MPG-since-last-fill preview as you type Last-fillup strip above the form + odometer prefill + one-tap +300 mi chip — see docs/user/odometer-prefill.md Offline queue that auto-syncs when signal returns iOS Shortcut integration (voice + deep-link) Stays on your network — backend talks to LubeLogger over the internal Docker network, not the public internet Screenshots Log Fuel History Maintenance Vehicles Settings Photo Preview Stats Quick start The fastest path to running quicklogger — standalone container, single host, no other services on the network. Five commands: git clone https://github.com/varunpan/quicklogger.git cd quicklogger cp compose.example.yml docker-compose.yml # Edit docker-compose.yml — set LUBELOGGER_URL to you","listed_at":1789166803,"md_updated_at":1789166803},{"full_name":"lasseh/taillight","url":"/r/lasseh/taillight","github":"https://github.com/lasseh/taillight","title":"taillight","tagline":"Real-time log viewer for syslog and application logs — filter, stream, and alert on events as they happen. Go backend with SSE streaming, Vue frontend, TimescaleDB storage, and pluggable notifications.","source":"trawl","tier":"found","license":"MIT","stars":34,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":559,"vulnerable":1},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a real-time syslog and application log viewer with a Go backend, Vue UI and live demo, \"Built with Claude Code as an implementation aid\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"taillight Lightweight, real-time log viewer for network operations teams. Stream netlog (network device syslog), srvlog (server syslog), and application logs to your browser with a single Go binary and PostgreSQL. Live Demo — try it without installing anything. More screenshots Why Taillight? Real-time first — events are pushed to the browser via Server-Sent Events the moment they hit the database. Per-client server-side filtering means you only receive what you asked for. Lightweight stack — PostgreSQL (TimescaleDB) + a single Go binary + rsyslog. No Elasticsearch, no JVM, no agent fleet to manage. Three log feeds — netlog (network device syslog), srvlog (server syslog), and applog (HTTP ingest) each have independent streams, filtering, retention, and dashboards. All three are always on; optional subsystems (notifications, AI analysis, LDAP, Netbox) are the real toggles. Multiple clients — a browser SPA in this repo, plus a terminal UI and SSH-served TUI in the companion taillight-tui repo — all consume the same SSE streams. Built-in alerting — Slack, webhook, email (SMTP), and ntfy channels. First matching event fires immediately; repeats during a silence window collapse into a single digest, with per-channel rate limiting, circuit breakers, and bounded delivery retry. Optional AI analysis — daily log briefings generated by a local Ollama LLM, surfacing anomalies and trends without sending data off-premise. Features Streaming & Filtering SSE push for netlog, srvlog, and","listed_at":1789166800,"md_updated_at":1789166800},{"full_name":"saphalpdyl/Aether","url":"/r/saphalpdyl/Aether","github":"https://github.com/saphalpdyl/Aether","title":"Aether","tagline":"Network Operations System from scratch(almost) with custom BNG routers. Built as a personal project to learn the ins and outs of ISP infrastructure emulated by containerlab.","source":"trawl","tier":"found","license":"MIT","stars":39,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an ISP network lab that emulates subscriber management with custom BNG routers, where \"the frontend, backend, and configuration were written by Claude\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"A full-stack ISP network lab that emulates IPoE IPv4 subscriber management end-to-end — from raw DHCP packet interception to real-time traffic dashboards. Built on an event-sourced architecture with RADIUS AAA, per-subscriber traffic shaping, and traffic simulation, all orchestrated via Containerlab. Table of Contents What is Aether? Why I Built This Architecture Overview Egress Packet Path How to Run Core Components BNG Control Plane Why Python? Session Creation through Sniffed DHCP Packets Data Plane DHCP Interception & Option 82 RADIUS Integration Per-Subscriber Traffic Shaping Event Bus (Redis Streams) Provisioning UI Network Topology (Containerlab) Simulator (How is traffic automatically generated?) Configuration-Driven Deployment Adding a Third BNG Infrastructure & Deployment Hosting Reverse Proxy & SSL CI/CD Pipeline Maintenance Mode Scheduled Redeployments Project Status & Roadmap Caveats & Known Limitations License Screenshots What is Aether? Aether is a multi-BNG config-driven ISP infrastructure lab built almost from scratch that emulates IPoE IPv4 subscriber management end-to-end. It supports IPoE/Ipv4 networks with RADIUS AAA, per-subscriber traffic shaping, traffic simulation and so on emulated by Containerlab. The entire lab is described in a single aether.config.yaml — BNG count, access node topology, subscriber subnets, DHCP pools, service IPs. Running make apply renders all the derived artifacts via Jinja2 templates: the Containerlab topol","listed_at":1789166796,"md_updated_at":1789166796},{"full_name":"goduni/unissh","url":"/r/goduni/unissh","github":"https://github.com/goduni/unissh","title":"unissh","tagline":"A cross-platform SSH client with end-to-end-encrypted vaults that sync through a server you host.","source":"trawl","tier":"found","license":"Apache-2.0","stars":124,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a cross-platform SSH client with end-to-end-encrypted vaults that can sync through a server you host, whose README signs off \"Built with Claude Code\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"UniSSH A cross-platform SSH client with end-to-end-encrypted vaults — local-first, and optionally synced through a server you host. Your keys, your hosts, your machine. The client is complete on its own — no account, no cloud, no server required. Run one only when you want the same vault on a second device or in a team, and even then it can't read a single byte of your data. ⬇ Download the client — macOS · Windows · Linux · Android · iOS Quick Start · Do I need the server? Contents Quick Start — get the client · self-host a server (optional) · admin panel Why UniSSH? Features Screenshots How it works · Identity model Components Installing unsigned builds — macOS · Windows · Linux · Android · iOS Configuration Security & Privacy Changelog — what changed, and whether it breaks your vault Build from source Contributing Community — supporting the project FAQ / Troubleshooting License Quick Start (A) is the whole install. Download the client and open it: the first run creates a local vault, and your hosts and keys are encrypted on that device. No account to create, no backend to stand up, and nothing of ours between you and the machines you connect to. (Do write down the Secret Key it shows you — it is what recovers the vault.) (B) and (C) exist only if you want the same vault on a second device or shared with a team. Everything the client does over SSH — sessions, SFTP, tunnels, fleet commands — behaves identically either way: SSH traffic goes straight from your device to you","listed_at":1789166794,"md_updated_at":1789166794},{"full_name":"adrhill/asdex","url":"/r/adrhill/asdex","github":"https://github.com/adrhill/asdex","title":"asdex","tagline":"Automatic Sparse Differentiation in JAX.","source":"trawl","tier":"found","license":"MIT","stars":47,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a JAX library that detects sparsity to compute Jacobians and Hessians with far fewer passes, whose README says \"This package is built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"asdex Automatic Sparse Differentiation in JAX. asdex exploits sparsity structure to efficiently materialize Jacobians and Hessians. It implements a custom Jaxpr interpreter that uses abstract interpretation to detect sparsity patterns from the computation graph, then uses graph coloring to minimize the number of AD passes needed. Installation pip install asdex Or with uv: uv add asdex Example import asdex import jax import jax.numpy as jnp def f(x): return (x[1:] - x[:-1]) ** 2 x_sample = jnp.zeros(50) # sample input for sparsity pattern detection jac_fn = jax.jit(asdex.jacobian(f, x_sample)) # ColoredPattern(49×50, nnz=98, sparsity=96.0%, JVP, 2 colors) # 2 JVPs (instead of 49 VJPs or 50 JVPs) # ⎡⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎤ ⎡⣿⎤ # ⎢⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⎥ → ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⠀⠀⎥ ⎢⣿⎥ # ⎢⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⎥ ⎢⣿⎥ # ⎣⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⎦ ⎣⠉⎦ for x in inputs: J = jac_fn(x) Instead of 49 VJPs or 50 JVPs, asdex computes the full sparse Jacobian with just 2 JVPs. Since sparsity detection and coloring can be expensive on large problems, we recommend saving and reusing colored patterns: import jax.numpy as jnp from asdex import jacobian_coloring from asdex import ColoredPattern, jacobian_from_","listed_at":1789166791,"md_updated_at":1789166791},{"full_name":"kcosr/termstation","url":"/r/kcosr/termstation","github":"https://github.com/kcosr/termstation","title":"termstation","tagline":"A web-based terminal session manager.","source":"trawl","tier":"found","license":"MIT","stars":63,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a web-based terminal session manager with layouts, scratch pads and an API, whose README warns \"100% of the code was generated by Claude and Codex.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"TermStation Pre-Alpha Warning This is a pre-alpha release. Use at your own risk! Review this README and the following documents before proceeding with installation. https://termstation.dev/faq CONTRIBUTING.md INSTALL.md docs/README.md (legacy, needs updates) docs/keyboard_shortcuts.md Overview TermStation is, at its core, a web-based terminal session manager that provides an API. It was created primarily to allow the developer access to his terminal sessions on multiple machines and to organize them in his preferred layout. Used regularly to manage agent CLIs, it does not have any knowledge of AI tools themselves. Integrations are built around templates where CLI commands and arguments are configured. It includes some features that extend further than basic terminal session management, including but not limited to: Template driven command definitions to drive GUI session creation forms Session and workspace-specific scratch pads Configurable code forges Launching commands in containers (e.g. agent CLIs) Generic PTY input API, used to provide agent-to-agent communication driven by prompting API for adding links to sessions and displaying them in the UI File generation with macros and includes, often used for .md file generation Reverse tunnel HTTP proxy for accessing web servers running in containers Multi-user support with RBAC and session sharing Requirements Backend Operating System: Ubuntu 22.04+ or Rocky Linux 10+ (other Linux distributions may work but are untested) Node","listed_at":1789166790,"md_updated_at":1789166790},{"full_name":"0xku/leetcode-compensation","url":"/r/0xku/leetcode-compensation","github":"https://github.com/0xku/leetcode-compensation","title":"leetcode-compensation","tagline":"Software engineering salaries (india) from leetcode compensation","source":"trawl","tier":"found","license":"MIT","stars":415,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":22,"vulnerable":3},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a filterable dashboard of Indian software engineer salaries parsed from Leetcode posts, where \"pretty much all of the html file has been generated by claude-opus-4.5\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Leetcode Compensation Software engineer salaries in India, parsed from Leetcode compensation posts and refreshed automatically. Leetcode Compensation fetches salary posts from Leetcode discussion forums, parses and normalizes them into structured data using LLMs, and presents everything in a filterable dashboard. Data stays fresh through automated GitHub Action PRs that sync new posts on a regular cadence. Getting Started Install uv from Standalone Installers or from PyPI: uv sync # Install all dependencies from pyproject.toml Updating Data The project uses LM Studio by default (LLM_PROVIDER=lm_studio) with the openai/gpt-oss-20b model for: Parsing salaries, years of experience (YOE), and other compensation details from posts Normalizing fields like companies, roles, and locations into structured format Run sync uv run leetcomp-sync Choose provider/model at runtime uv run leetcomp-sync --provider llama_server --model unsloth/Qwen3.5-9B-GGUF Supported providers: lm_studio (default) llama_server (also supports alias llama-server) github_models (requires GITHUB_TOKEN) zai (requires ZAI_API_KEY) Optional env overrides: LLM_PROVIDER LLM_MODEL LLM_BASE_URL Example: llama.cpp server /opt/homebrew/bin/llama-server \\ --hf-repo unsloth/Qwen3.5-9B-GGUF \\ --hf-file Qwen3.5-9B-Q4_K_M.gguf \\ --port 5000 \\ -c 65536 uv run leetcomp-sync --provider llama-server --model unsloth/Qwen3.5-9B-GGUF LLM Assistance I've written all the data parsing logic in python by hand. Most of the prompts have be","listed_at":1789166785,"md_updated_at":1789166785},{"full_name":"Nyein-Sat-Thaw/myanmar-quiz-game","url":"/r/Nyein-Sat-Thaw/myanmar-quiz-game","github":"https://github.com/Nyein-Sat-Thaw/myanmar-quiz-game","title":"myanmar-quiz-game","tagline":"🇲🇲 Myanmar Quiz Game — Test your knowledge about Myanmar history, culture, geography, and food! Built with Claude Code.","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a browser quiz game about Myanmar history, culture, geography and food, whose description says it was \"Built with Claude Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🇲🇲 Myanmar Quiz Game Ever wondered how much you really know about Myanmar? I built this little quiz game to test that. It's a simple web app — no install, no sign-up, just open index.html in your browser and start playing. You pick a category (history, culture, geography, or food), and it gives you 10 random questions. Get instant feedback on each answer, and see your final score at the end. I made this because I wanted something fun and educational at the same time. There are over 40 questions covering everything from ancient kingdoms to mohinga. How to play Clone the repo and open it in your browser: git clone https://github.com/Nyein-Sat-Thaw/myanmar-quiz-game.git Then just open index.html. That's it. What's inside 4 categories — History, Culture, Geography, Food (or mix them all) 40+ questions with 4 choices each Score tracking as you go Review screen at the end so you can learn from mistakes The categories 🏛️ History — kings, independence, colonial era, modern events 🎭 Culture — festivals, traditions, thanaka, puppet shows 🗺️ Geography — rivers, lakes, mountains, capitals 🍜 Food — mohinga, lahpet, ohn no khao swè, and more How it's built Plain HTML, CSS, and JavaScript. Nothing fancy. No frameworks, no build steps. I used Claude Code to help me write the code, generate quiz questions, and set up the project structure. Adding your own questions Want to add more? Open questions.js and follow the same format: { q: \"Your question here?\", options: [\"A\", \"B\", \"C\", \"D\"], ","listed_at":1789166783,"md_updated_at":1789166783},{"full_name":"micky-gee/intervals-glances","url":"/r/micky-gee/intervals-glances","github":"https://github.com/micky-gee/intervals-glances","title":"intervals-glances","tagline":"A vibe coded Garmin ConnectIQ app to display intervals.icu data","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Monkey C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Garmin watch widget that shows intervals.icu training data, described by its owner as \"A vibe coded Garmin ConnectIQ app to display intervals.icu data\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Intervals — an intervals.icu widget for Garmin watches A Connect IQ widget with a glance that brings your intervals.icu fitness data to round-screen Garmin watches. Built for and tuned on the Fenix 8 Pro MicroLED (fenix8pro47mm, 454×454, API 6.0); supports 66 device IDs across the Fenix 7/8/9/E, Epix 2, Forerunner, Venu, MARQ 2, Descent, Approach and Instinct 3 AMOLED families (layouts scale with screen size; MIP models render the chart palette with fewer colors). Glance Form Load (polar) HRV ring Ramp rate ring Recovery tiles More: eFTP · steps · sleep · body · fuel Features Glance — a compact banded CTL/ATL chart (the form-zone colors at a glance) with the current form value and data age, right in the widget carousel. Form page — big TSB value with zone label, plus Fitness / Fatigue / Ramp rate. Polar load chart — CTL and ATL wrapped around the bezel over the time-varying form-zone bands (transition / fresh / grey / optimal / high-risk), computed per angle step from interpolated CTL. The display range auto-fits the data and band envelope; labeled value arcs and 30-day spokes keep it readable. Every graph page can render round (radial) or rectangular, per page. Four configurable graph pages — each picks the load chart or one of 17 metrics: HRV (rMSSD/SDNN), resting/sleeping HR, ramp rate, eFTP, weight, body fat, VO2max, sleep score/hours, readiness, steps, SpO2, respiration, Baevsky stress index, calories. Radial bar rings (newest day highlighted) or auto-scaled line charts ","listed_at":1789166782,"md_updated_at":1789166782},{"full_name":"nullstyle/capnp-zig","url":"/r/nullstyle/capnp-zig","github":"https://github.com/nullstyle/capnp-zig","title":"capnp-zig","tagline":"Vibe-coded attempt at building a capnproto implementation in zig","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Zig","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a pure Zig implementation of Cap'n Proto serialization and RPC, described by its owner as a \"Vibe-coded attempt at building a capnproto implementation in zig\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"capnpc-zig A pure Zig implementation of Cap'n Proto -- a serialization framework and RPC system. Includes a compiler plugin (capnpc-zig), a message serialization library, and an RPC runtime built on std.Io with a concurrent read/write transport. Targets Zig 0.17-dev. Status (v0.18.0): serialization, codegen, the capnpc-zig plugin, and the two-party RPC core are Stable on a frozen, CI-gated public surface (docs/api-snapshot.txt). The L3/L4 three-party arc, the reflected-cap resolver, QUIC, persistence vat-restore, events, binary schema reflection, and the demoted transport/ctor variants remain Experimental and may change at any 0.x minor bump. Pre-1.0 — pin an exact version. See docs/supported-surface.md for the full contract and docs/stability.md for the per-module matrix. Features Pure Zig Implementation: No C++ dependencies, targets Zig 0.17-dev Full Serialization Support: Complete Cap'n Proto wire format including packed encoding and far pointers Zero-Copy Deserialization: Readers work directly with message bytes Builder Pattern: Ergonomic API for constructing messages Schema-Driven Code Generation: Generates idiomatic Zig Reader/Builder types from .capnp schemas Executable Type Fidelity: Brand-aware schema validation/canonicalization and finite typed generic views without removing erased APIs Binary Schema Reflection (Experimental): Embedded schema nodes, type and field lookup, and dynamic readers/builders with generic bindings and schema evolution RPC Runtime: Cap'n Prot","listed_at":1789166780,"md_updated_at":1789166780},{"full_name":"arongate/ubuntu-miracast-client","url":"/r/arongate/ubuntu-miracast-client","github":"https://github.com/arongate/ubuntu-miracast-client","title":"ubuntu-miracast-client","tagline":"Open-Source  vibe-coded ubuntu miracast client","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an Ubuntu desktop app for casting your screen to Miracast devices, described by its owner as a \"vibe-coded ubuntu miracast client\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Ubuntu Miracast Client A desktop application for Ubuntu that enables screen and application casting to Miracast-compatible devices. ⚠️ Unstable Phase (0.x) This project is in its initial development phase (0.x.y). Per SemVer §4, the public API is not yet stable — any release may introduce breaking changes. Pin your dependency to an exact version if you rely on this package. Features Cast your entire screen or specific application windows Discover Miracast receivers on your network via Wi-Fi Direct WFD protocol compliant — full RTSP M1-M7 session negotiation (Wi-Fi Display spec v2.3) Real-time casting session statistics (bitrate, data transferred, dropped frames) Session history tracking with detailed statistics Optional systemd user service mode Modern GTK 4 + libadwaita interface Configurable video quality, frame rate, and audio streaming Prerequisites Ubuntu 24.04 LTS (or compatible Linux distribution) Python 3.10 or higher Network connection (5GHz Wi-Fi recommended for optimal performance) System Dependencies sudo apt install python3-gi python3-cairo python3-gst-1.0 \\ gir1.2-gtk-4.0 gir1.2-adw-1 \\ gir1.2-gstreamer-1.0 gir1.2-gst-plugins-base-1.0 \\ gstreamer1.0-tools gstreamer1.0-plugins-good gstreamer1.0-plugins-ugly \\ wpasupplicant x11-utils Installation From Debian Package sudo apt install ./ubuntu-miracast-client_0.0.1_amd64.deb From Source git clone https://github.com/arongate/ubuntu-miracast-client.git cd ubuntu-miracast-client # Install uv (if not already installed) ","listed_at":1789166779,"md_updated_at":1789166779},{"full_name":"Fabian2315/Multi-Bot","url":"/r/Fabian2315/Multi-Bot","github":"https://github.com/Fabian2315/Multi-Bot","title":"Multi-Bot","tagline":"Controll Multiple bots through a modern web-interface. Made using mineflayer","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["bot"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":360,"vulnerable":13},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a web dashboard for controlling multiple Mineflayer Minecraft bots, whose owner says \"This project is largely vibecoded (including this readme)\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Multi-Bot A Mineflayer-based Minecraft bot controller with a local web dashboard, now supporting multiple bots, command targeting, groups, and a single switchable Prismarine viewer. Important This project is largely vibecoded (including this readme), because im a bad dev and to be honest as a student dont want to spend my free time learning how to code. Please expect some type of unexpected behavieour and bugs. Chat Notice If you want bots to reply to in-game chat commands, make sure your server allows unsigned chat as needed for your setup. For many offline/local setups, this means setting the following in server.properties: enforce-secure-profile=false Project Overview The app starts one starter bot and hosts a local dashboard. From the dashboard you can: Add more bots (single or batch naming) Control one bot, all bots, or a group Create and edit groups for easier command targeting Configure starter auth mode (offline, Microslop, token) Switch the single Prismarine viewer to any bot Features Multi-bot runtime manager Per-target command execution: Single bot All bots Group Staged command execution Multiple queue save slots (quick save/load/run) Group management (create, edit, delete) Bot automation capabilities: Movement and following PvP and guard mode Self-defense retaliation Block collection and mining loop Craft planning and table-assisted crafting Furnace/smoker/blast-furnace smelting Auto-eat and manual eat Inventory emptying to nearest chest Context-aware command inpu","listed_at":1789166777,"md_updated_at":1789166777},{"full_name":"briandenicola/Aurearia","url":"/r/briandenicola/Aurearia","github":"https://github.com/briandenicola/Aurearia","title":"Aurearia","tagline":"Self-hosted app for cataloging, organizing, and sharing your coin collection","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["web-app"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":917,"vulnerable":2},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted web app for cataloging and sharing a personal coin collection, whose README says \"This application is 100% vibe coded.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Aurearia A self-hosted Progressive Web App for cataloging, analyzing, and sharing a personal coin collection. Note: This application is 100% vibe coded. It exists for the author to learn and experiment with GitHub Copilot CLI. What is this? Aurearia is a self-hosted, full-featured Progressive Web App for managing a personal coin collection end-to-end, with an emphasis on ancient and historical coins. Catalog, value, analyze, organize, and share your coins with powerful AI-assisted tools, offline access, and community features — all on your own server. For collectors who want: 🏛️ Detailed numismatic records with provenance, images, and structured references 🤖 AI-powered analysis of coin photos (Claude or self-hosted Ollama) 🔍 Smart discovery tools for finding, tracking, and monitoring coins 📊 Rich portfolio analytics with value trends and gap analysis 🤝 Social features to follow collectors and share curated collections 📱 Install as a PWA; offline read access on mobile and desktop 🔒 Complete control of data; runs entirely on your infrastructure Built for a single collector plus invited friends. Not a marketplace, SaaS, or reference database replacement. If you expose it outside a trusted home network, follow the public-facing hardening checklist in docs/deployment.md: TLS, reverse-proxy security headers, trusted proxy IP handling, invite/closed registration, private agent networking, backups, and monitoring are required. Product vision, personas, and detailed requirement","listed_at":1789166775,"md_updated_at":1789166775},{"full_name":"travisbreaks/kimi-agent-mandala","url":"/r/travisbreaks/kimi-agent-mandala","github":"https://github.com/travisbreaks/kimi-agent-mandala","title":"kimi-agent-mandala","tagline":"Interactive 3D Sri Yantra mandala with 50K+ particles. Scroll-driven sacred geometry that builds, breathes, and explodes. React Three Fiber + custom GLSL shaders.","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["art"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an interactive 3D Sri Yantra mandala of 50K particles driven by scroll, whose README says it was \"Built with Claude Code and inspired by sacred geometry traditions.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Kimi Agent Mandala An interactive 3D visualization of a Sri Yantra mandala that responds to scroll, hover, and click interactions. Built with React Three Fiber and custom WebGL shaders. Features Two-Phase Scroll Animation Phase 1 (0-50% scroll): Build - Particles fly in from all directions to form the sacred Sri Yantra geometry Phase 2 (50-100% scroll): Explode - The mandala explodes outward with a spiral motion while a fractal tunnel emerges in the background Interactive Elements Hover: Tilt the mandala based on pointer position Click: Create a pulse effect that radiates through the particles Scroll Velocity: Particle spin speed and visual effects respond to scroll speed Accessibility & Performance Adaptive Quality: Automatically detects GPU capabilities and adjusts particle count High-end: 50,000 particles Medium: 25,000 particles Low-end/integrated GPUs: 10,000 particles Reduced Motion: Respects prefers-reduced-motion media query, reducing particle count for accessibility Error Boundary: Graceful error handling with fallback UI Tech Stack Core React 19.2 - UI framework TypeScript 5.9 - Type safety Vite 7.2 - Build tool & dev server 3D Graphics Three.js 0.182 - WebGL 3D library @react-three/fiber - React renderer for Three.js @react-three/drei - Helper components for R3F Styling Tailwind CSS - Utility-first CSS framework Custom Shaders - GLSL vertex & fragment shaders for procedural effects Getting Started Prerequisites Node.js 18+ npm or yarn Installation # Clo","listed_at":1789166770,"md_updated_at":1789166770},{"full_name":"FaranAiki/alkyl","url":"/r/FaranAiki/alkyl","github":"https://github.com/FaranAiki/alkyl","title":"alkyl","tagline":"Experimental programming language written in \"pure\" C with QBE and LLVM as the backend. ARCUY \"Always Reasonably Choose Your Way\"","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an experimental programming language written in C with LLVM and QBE backends, which its author says was \"mostly vibe coded\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Alkyl Programming Language A simple programming language written using C as a frontend and backends, such as LLVM-IR (mainly), QBE (experimental, but well done), MLIR (very experimental), and Cranelift (mismatch between standard library), and self-backend (to be implemented), with the philosophy of ARCUY: Always Reasonably Choose Your Way. Alkyl is named after aikil (aiki language: sorry, kind of NPD here as this is a personal project) and alkyl in chemistry which is a highly reactive and foundational. This means alkyl is a very modularized language that can acts as others. The file extension for Alkyl source code is .kyl, a sweet suggestion made by my beloved girlfriend (saysay/kupkup), whereas for the header file it is .hky because it sounds... hacky. Purpose Alkyl is a general multipurpose language that is used for experimenting with lexer, parser, and code generation. Its principle is ARCUY: \"Always Reasonably Choose Your Way\" which means the keyword \"reason\" is embedded in its lexer and parser. Moreover, Alkyl is similar to C, Dart, Zig, and some other programming languages. Alkyl does not try to compete with other languages, even languages like Racket or Lisp. It is used for personal project and to understand how old compilers and modern compilers work (duh, mostly vibe coded so I didn't 101% rtfm). This is why Alkyl is not a revolution in programming language, it is just a side hobby project that may be useful to me and others (99% not useful). Pipeline Supported Backe","listed_at":1789166769,"md_updated_at":1789166769},{"full_name":"LckHot/MathMD","url":"/r/LckHot/MathMD","github":"https://github.com/LckHot/MathMD","title":"MathMD","tagline":"Definitely a vibe coding result. Minimal Android markdown reader for mathematicians — renders $...$, $$...$$, \\(...\\), \\[...\\] correctly (KaTeX, offline).","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a minimal Android markdown reader that renders TeX math for mathematicians, whose owner says \"every line of this repository was written by an AI agent\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"MathMD In the AI-for-research era, GPT and friends answer math questions in Markdown. Android had no clean reader that renders all four common TeX/LaTeX math delimiter styles — so this simple app now exists. A minimal Android markdown reader (with light editing) built for mathematicians. It renders TeX/LaTeX math correctly in all four common delimiter styles: Inline: $x_i$ and \\(u \\in C^k\\) Display: $$...$$ and \\[ ... \\] Why another markdown app? Most Android markdown editors either render no math at all, or break on real mathematical documents: the markdown parser runs first and eats \\(...\\) / \\[...\\] backslashes (CommonMark escape rules) and mangles _-containing formulas like $\\bar{\\mu}_{n}$ into emphasis spans before any math renderer sees them. MathMD extracts math segments before markdown parsing (a protection/placeholder pass), so all four delimiter styles survive intact. The pass is written in TypeScript and covered by a regression suite (including $\\{x\\}$, inline pairs, $W^{1,p}_0(\\Omega)$, and code spans/fences that must stay literal). Features Four math delimiter styles, rendered by bundled KaTeX (fully offline) Edit / Preview flip toggle; configurable startup mode In-document search with match highlighting and next/previous navigation, working in both the editor and the rendered preview New blank document; open via system file picker or \"Open with…\" from other apps; save in place or as a copy Light/dark/system theme; per-pane font family and editor font size Fixed ","listed_at":1789166749,"md_updated_at":1789166749},{"full_name":"Titanean/Tadmor","url":"/r/Titanean/Tadmor","github":"https://github.com/Titanean/Tadmor","title":"Tadmor","tagline":"An exoplanet database and tracker with procedural visualisations.","source":"trawl","tier":"found","license":"Apache-2.0","stars":4,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Tadmor, an Android exoplanet catalog that renders each planet as a procedural 3D globe; its README says \"Tadmor was 100% vibe coded\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Tadmor An exoplanet database and tracker with procedural visualisations. Tadmor pulls the latest data from the NASA Exoplanet Archive — currently 6,000+ confirmed planets plus TESS / Kepler / K2 candidates — and renders each one as a 3D globe built procedurally from its measured parameters. The same physical inputs that drive the classifier (mass, radius, temperature, host spectral type, atmospheric composition, orbital geometry) drive the look of the planet, its host star, and the orbital diagram. Nothing is a photograph. Features Catalog — browse the full planet list with search, multi-axis filters (composition, temperature, mass class, host spectral type, discovery method, data completeness), and sort. Per-planet 2D icons sketch the body at a glance. Candidates — separate tabs for confirmed planets, TESS / Kepler / K2 candidates, and false positives, with dispositions kept up to date by the sync. Bookmarks — save any planet; the catalog row gets a gold-edge highlight and a saved-only filter is one tap away. The app snapshots the planet's parameters at bookmark time. Update tracking — when a bookmarked planet's catalog parameters change after a sync (a refined mass measurement, a new temperature, a disposition flip), the card shows an \"n updated\" badge and the updated fields are diffed against the snapshot. Background sync — WorkManager periodically pulls the latest data from the NASA Exoplanet Archive so the local catalog stays current without manual refresh. System view —","listed_at":1789166747,"md_updated_at":1789166747},{"full_name":"cursork/gritt","url":"/r/cursork/gritt","github":"https://github.com/cursork/gritt","title":"gritt","tagline":"Go! Ride!","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a terminal IDE for Dyalog APL written in Go, whose owner says \"This is a Claude Code project; unashamedly so.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"gritt A terminal IDE for Dyalog APL. Pronounced like \"grit\" (G from Go + German \"Ritt\" = ride). WARNING This is an alpha level project. I will promote it to 0.1.0 when I think it's ready for broad use. Here be dragons. I use it daily, but mostly for specific tasks. I revert to Ride otherwise, since it's complete. Features Full TUI with floating panes for editors, tracer, debug info APL input: backtick prefix (`i → ⍳), symbol search, APLcart integration Debugging: breakpoints, stepping (into/over/out), stack trace, variables pane, edit while debugging (very much a 'maybe' - don't trust it) Command palette for quick access to all commands Connection resilience - stays alive on disconnect, allows reconnect Single-expression and stdin modes for scripting Link integration for source-controlled APL projects Tracer with stack navigation (single pane, not overlapping windows) Edit a function/namespace/array in your preferred $EDITOR (C-] e) — saves back via SaveChanges on exit See example-test-report.html or example-test-report.txt for a walkthrough of features (snapshots from automated tests). Installation Download Grab a binary from Releases: gritt-darwin-arm64 - macOS Apple Silicon gritt-darwin-amd64 - macOS Intel gritt-linux-arm64 - Linux ARM64 gritt-linux-amd64 - Linux x86_64 gritt-windows-amd64.exe - Windows x86_64 gritt-windows-arm64.exe - Windows ARM64 Windows note: SmartScreen will block the download (click \"Keep\" in Edge, or \"Keep anyway\" in the dialog). Defender may also f","listed_at":1789166746,"md_updated_at":1789166746},{"full_name":"Batushn/deskflick","url":"/r/Batushn/deskflick","github":"https://github.com/Batushn/deskflick","title":"deskflick","tagline":"Mouse gestures for KDE Plasma: hold a mouse button and flick to switch virtual desktops or snap windows. Wayland & X11.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["productivity"],"built_with":["claude"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: deskflick, hold-and-flick mouse gestures for switching desktops and snapping windows in KDE Plasma, badged by its owner \"Vibe coded with Claude\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"deskflick Mouse gestures for KDE Plasma. Hold a mouse button, flick, and glide between virtual desktops — or snap the window you're looking at. Your mouse has a spare button — the \"back\" thumb button on most mice. deskflick turns it into a whole navigation layer: Gesture Default Hold + flick ← → ↑ ↓ Switch virtual desktop that way Meta + hold + flick Snap the focused window that way (like Meta + arrows) Hold + flick ↖ ↗ ↙ ↘ Move one desktop on both axes at once Meta + hold + flick ↖ ↗ ↙ ↘ Snap the window into that corner Tap Show desktop Double tap Maximise window Press and hold Show all windows (Present Windows) Keep holding and keep pushing to fly across your whole desktop grid. Every one of those is rebindable to any KWin shortcut or shell command, and a plain tap can simply pass through as the button's original click. Requirements KDE Plasma 5 or 6, on Wayland or X11 Python 3.11+ and python-evdev pyside6 for the settings window (optional — the daemon runs without it) Arch / CachyOS / Manjaro: sudo pacman -S python-evdev pyside6 Debian / Ubuntu: sudo apt install python3-evdev python3-pyside6 Fedora: sudo dnf install python3-evdev python3-pyside6 Install git clone https://github.com/Batushn/deskflick && cd deskflick && ./install.sh That copies deskflick and deskflick-ui into /usr/local/bin, installs a udev rule granting your session read access to mouse devices, and enables the deskflick systemd user service. It asks for sudo once. Then hold your mouse's bac","listed_at":1789166745,"md_updated_at":1789166745},{"full_name":"claygorman/aw5d-linux","url":"/r/claygorman/aw5d-linux","github":"https://github.com/claygorman/aw5d-linux","title":"aw5d-linux","tagline":"Drive the iBUYPOWER AW5/AW5D AIO cooler's round LCD from Linux (native HID, no vendor software)","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a script that drives the iBUYPOWER AW5D AIO cooler's round LCD from Linux; the README says \"the code and documentation in this repository were written by AI\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"aw5d-linux Drive the iBUYPOWER AW5 / AW5D 360 mm AIO cooler's round LCD from Linux — no Windows, no vendor software. WarningAI-authored · use at your own risk. The code and documentation in this repository were written by Claude (Anthropic's Claude Code) working with the author, from black-box reverse-engineering of the hardware (details: Authorship & AI disclosure and How this was figured out). It ships no vendor code. It's a small, dependency-free, auditable script that only writes the documented status report to a hidraw device — but it comes with absolutely no warranty; you run it entirely at your own risk (see LICENSE). Read it before you run it. The AW5D's stock \"digital gauge\" (big CPU-temperature number, spinning fan, coloured arc + bars) is drawn and animated by the cooler's own firmware. All the host has to do is push one 64-byte USB-HID report per second with the live CPU stats. This project does exactly that, reading temperature / usage / clock straight from the Linux kernel — so the screen just works, natively. Status: working on an iBUYPOWER unit (3402:0407, \"CoolerMaster\" variant) on a Ryzen 7 7800X3D / Bazzite box. The three real readouts — temperature, usage, clock — are exact. Contributions/other variants welcome. Why this exists There is (was) no Linux support for this cooler in liquidctl, CoolerControl, or any community project — the vendor only ships a Windows app (HYTE Nexus). The cooler itself, though, is a plain USB-HID device that Linux can talk t","listed_at":1789166743,"md_updated_at":1789166743},{"full_name":"girving/ltop","url":"/r/girving/ltop","github":"https://github.com/girving/ltop","title":"ltop","tagline":"Minimal process monitor similar to top","source":"trawl","tier":"found","license":"BSD-2-Clause","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: ltop, a minimal tree-style process monitor in Rust with per-process GPU usage, whose README says \"All code was written by Claude Opus 4.6 through 4.8\". It carries the BSD-2-Clause license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"ltop 🌳 Minimal process monitor (top replacement) in Rust. Designed for monitoring builds (Lean/Lake, C++, etc.), GPU experiments, and similar workloads. Screenshot: Features: Tree-structured process display Per-process GPU usage column (NVIDIA on Linux, AGX on macOS) with no cover library linked — raw /dev/nvidiactl ioctls and raw IOKit MIG CPU and memory threshold filtering Interpreter-aware: shows script names for python, ruby, node, etc. Color-coded by CPU (blue) and memory (red) pressure Works on Linux and macOS Footprint The shipped builds run libc-free static (Linux) or single-LC_LOAD_DYLIB to libSystem only (mac); GPU monitoring is built in without dragging the cover library into the link. Target Binary Startup RSS Stable RSS Linux x86_64 ≤ 23.7 KB ~44 KB ~44 KB Linux aarch64 ≤ 22.0 KB ~44 KB ~44 KB macOS arm64 † ≤ 32.6 KB ~944 KB ≤ 224 KB † macOS phys_footprint peaks transiently at ~944 KB during dyld init while libSystem subdylibs load and libmalloc reserves zone arenas; mac_sys::unmap_idle_state reclaims that state at the top of main before the first tick fires, dropping to the 224 KB steady-state. Of those 224 KB, ~160 KB is per-process page-table backing the dyld shared cache, which mach_vm_deallocate can't actually clear (kernel ignores deallocation of shared submap entries — see CLAUDE.md \"mach_vm_deallocate is a placebo on dyld-shared-cache submap entries\"). Linux has no equivalent transient peak: the libc-free static build goes from kernel exec straight to _s","listed_at":1789166742,"md_updated_at":1789166742},{"full_name":"derfsss/rust-for-amigaos4","url":"/r/derfsss/rust-for-amigaos4","github":"https://github.com/derfsss/rust-for-amigaos4","title":"rust-for-amigaos4","tagline":"Write native AmigaOS 4.1 applications and device drivers in Rust","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Rust for native AmigaOS 4.1 apps and drivers; \"This project (code, bindings, build infrastructure, and documentation) was built using Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Rust for AmigaOS 4 Write native AmigaOS 4.1 applications, device drivers, and shared libraries in Rust. Status: Beta — 31 safe wrapper modules, 129 SDK interface bindings, ~365 tests, 3 build modes. Tested on QEMU (-M amigaone) and real X5000 hardware. Made with AI — This project (code, bindings, build infrastructure, and documentation) was built using Claude Code by Anthropic. Features Full no_std Rust — core and alloc crates compile to PowerPC Vec, String, format!, Box — heap allocation via global allocator Three build modes — application (clib4), driver (ExecAllocator), shared library (Resident + interface vectors) 31 safe wrapper modules — GUI (ReAction), menus, ASL file requesters, networking (TCP/DNS/HTTP/HTTPS), async runtime, timer device, clipboard (IFFParse), DOS, file I/O, threads, and more 129 AmigaOS SDK interface bindings — Exec, DOS, Intuition, Graphics, Timer, IFFParse, and 123 more, all feature-gated Direct vtable dispatch — call any interface method from Rust via #[repr(C)] structs (no overhead) RAII everywhere — AmigaWindow, AmigaTimer, AmigaLock, AmigaVec, TcpStream, PubScreen, and 14 more auto-cleanup on drop Checked OS strings — amstr!(\"text\") builds null-terminated strings at compile time; every wrapper validates at the FFI boundary ReAction GUI — LayoutBuilder DSL, event_loop, button/string/checkbox/integer/slider/chooser/listbrowser gadgets Menus — MenuBuilder DSL over the OS4 menuclass, with MM_NEXTSELECT pick decoding ASL file requester — Open/Save/","listed_at":1789166740,"md_updated_at":1789166740},{"full_name":"BZLZHH/LavaFlow","url":"/r/BZLZHH/LavaFlow","github":"https://github.com/BZLZHH/LavaFlow","title":"LavaFlow","tagline":"An experimental Vulkan 1.1 graphics backend for Minecraft's Blaze3D API.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"Java","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: LavaFlow, an experimental Vulkan graphics backend for Minecraft, whose README says it is \"100% generated by AI under human direction\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"LavaFlow Caution100% AI-GENERATED PROJECT LavaFlow, including its source code and documentation, is 100% generated by AI under human direction. Treat the implementation as experimental: audit the code and test it on your own hardware before relying on it. LavaFlow is an experimental Vulkan 1.1 graphics backend for Minecraft's Blaze3D API. It replaces the exercised Blaze3D rendering path with LavaFlow-owned LWJGL Vulkan implementations and never calls OpenGL. Compatibility Component Target Minecraft 26.2 Mod loader NeoForge 26.2.0.32-beta or newer 26.2 build Minecraft runtime Java 25 Graphics API Vulkan 1.1 Desktop smoke renderer Java 21 bytecode The backend includes compatibility paths for Vulkan 1.1 devices without dynamic rendering, synchronization2, push descriptors, multi-draw indirect, non-solid fill mode, or vertex attribute divisors. Both desktop and ARM64 Android devices are in scope. Actual driver behavior and performance vary by GPU. Current Features LavaFlow-owned Vulkan instance, presentation device, queues, and swapchain Blaze3D textures, buffers, samplers, pipelines, render passes, and command encoding Dynamic-rendering and legacy-render-pass paths Push-descriptor and descriptor-set paths Explicit resource lifetime and queue synchronization Vulkan-native texture transfers, blits, clears, and presentation Resize and swapchain recreation handling Final presentation blit with the target Y coordinate reversed Sodium 0.9.1 compatibility: Sodium's Vulkan terrain path ","listed_at":1789166739,"md_updated_at":1789166739},{"full_name":"pablojimenezmateo/microide","url":"/r/pablojimenezmateo/microide","github":"https://github.com/pablojimenezmateo/microide","title":"microide","tagline":"A 100% vibecoded native, low-footprint, keyboard-first C++/SDL3 desktop IDE for Linux — built-in editor, diff/merge, git, search, terminal, and debugger.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["devtools","cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: microide, a native keyboard-first C++/SDL3 desktop IDE for Linux with diff, git and a debugger, which its README calls \"100% vibecoded\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"microide A 100% vibecoded, native, low-footprint C++/SDL3 desktop IDE focused on built-in editor, diff, merge, git, search, and terminal workflows. Single-window, keyboard-first, and does not require a GPU (it uses one when available to speed up text, and runs fully on a software renderer otherwise). Editor, file tree, syntax highlighting, git blame, debugger pane, and terminal — captured straight from the running app. See the full screenshot gallery and hero demo video on the project site. For the authoritative in-scope / non-goal list see openspec/specs/product-vision/spec.md. Status: stable. Tagged v2.11.1 (see CHANGELOG) and actively developed. Release binaries are GPG-signed — verify them per Verifying releases. No third-party comparative benchmarks yet. Read Known Limitations and Security & Trust Model before using on a real project. Verified, not just written. 121 test files and 10 fuzz targets back the build, gated by ASAN / UBSAN / TSAN sanitizer runs, committed startup/typing/scroll/diff/search performance baselines, and architecture-invariant lints enforced on every test run. Start Here What microide is Release status Current UI preview What works today Agent control channel Build or package locally Known limitations Plugin trust warning Performance methodology summary Deeper docs Contributing About microide is 100% vibecoded: every source file, test, and document in this repository was written by AI coding agents (a mix of Claude, GPT, and other tools) under h","listed_at":1789166738,"md_updated_at":1789166738},{"full_name":"N3rmis/Godot-Team-Create","url":"/r/N3rmis/Godot-Team-Create","github":"https://github.com/N3rmis/Godot-Team-Create","title":"Godot-Team-Create","tagline":"Godot Team Create plugin that lets you connect with your friends and live edit the game together!","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"GDScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Godot 4 plugin for live collaborative game editing with friends, whose README admits \"it's like really vibecoded so feel free to throw tomato at me\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"-Artwork by: cosogamer97 WarningThis plugin is still in development! bugs can happen idk report them to me i guess. and it's like really vibecoded so feel free to throw tomato at me, atleast i got it to work somewhat so thats nice 👍 This is a Godot 4.4+ plugin that lets you and your friends build the game together in real-time. What you need Godot 4.4 or newer. A decent internet connection between everyone. OPTIONAL: Hosted server, it's just more stable for some reason How to get it running Download the latest release. Drop files into your project's addons/ folder. Open Godot, go to Project Settings -> Plugins. Check the box to turn it on. How to connect LAN Connection Host: Open the Team Create dock. Click Host LAN. This will start a server on your local machine (default port 25567). Give your local IP address to your friends on the same network. User: Open the Team Create dock. Enter the Host's local IP address in the text box. Click Join. 📖 Check out the Team Create Wiki for detailed guides on connecting via Hamachi / Radmin VPN, dedicated servers, commands, and features! Wanna help? Any testing is already a huge help lolololo Breaking things: If you find a bug, open an issue and let me know exactly how you broke it so I can try to fix it Fixing things: If you want to write some code feel free to do pull requests. Just open a discussion first if you wanna make something big","listed_at":1789166736,"md_updated_at":1789166736},{"full_name":"nicedreamzapp/Family-Planner","url":"/r/nicedreamzapp/Family-Planner","github":"https://github.com/nicedreamzapp/Family-Planner","title":"Family-Planner","tagline":"Self-hosted family command center for a spare iPad — voice control, OCR document scanning, meal planning, chore tracking, and seasonal themes. No subscriptions, no cloud.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["web-app"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted family command center for a spare iPad with calendar, meals, chores and voice control, whose README says \"Built with Claude Code assistance.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Family Planner Dashboard A Skylight Calendar-inspired family command center with voice control, OCR document scanning, AI assistant, and dynamic seasonal themes. Built this because I wanted a fully customizable family hub that I control. Runs on a spare iPad and does everything - voice commands, document scanning, meal planning, chore tracking with celebrations, and ambient video screensavers. No subscriptions, no cloud dependency. What It Does ┌─────────────────────────────────────────────────────────────────┐ │ 🎙️ VOICE │ 📸 OCR SCAN │ 🤖 AI CHAT │ 📺 SCREENSAVER │ ├─────────────────────────────────────────────────────────────────┤ │ │ │ 📅 CALENDAR │ 🍽️ MEAL PLANNER │ │ ├─ Events │ ├─ Weekly meals │ │ ├─ Holidays │ ├─ School lunch menus │ │ ├─ Moon phases │ └─ Voice: \"Tacos for Friday\" │ │ ├─ Weather forecast │ │ │ └─ Seasonal themes │ 🛒 SHOPPING LIST │ │ │ └─ Voice: \"Add milk\" │ │ ⭐ KIDS CHORES │ │ │ ├─ Task tracking │ 🎬 MONTHLY SCREENSAVERS │ │ ├─ Celebrations! │ ├─ Seasonal video backgrounds │ │ └─ No points needed │ └─ Ocean idle screensaver │ │ │ │ └─────────────────────────────────────────────────────────────────┘ Features 🎙️ Voice Commands (Works Everywhere) Just talk to it. Uses Web Speech API with OpenAI Whisper fallback for accuracy. \"Add dentist appointment Tuesday at 3pm\" \"Emma brushed her teeth\" \"Plan spaghetti for Wednesday\" \"Add eggs to the shopping list\" The voice bar is everywhere - calendar, meals, chores, shopping. Speak naturally. 📸 OCR Document S","listed_at":1789166735,"md_updated_at":1789166735},{"full_name":"LeZergan/pvz2vita","url":"/r/LeZergan/pvz2vita","github":"https://github.com/LeZergan/pvz2vita","title":"pvz2vita","tagline":"Unofficial PS Vita loader for the Android 4.5.2 ROW build of Plants vs. Zombies 2.","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an unofficial PS Vita loader for the Android build of Plants vs. Zombies 2, whose README warns it was \"Fully vibecoded, directed and tested by LeZergan.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Caution WARNING: THIS PORT WAS WRITTEN WITH AI The port-specific code and this documentation were generated with OpenAI Codex. Fully vibecoded, directed and tested by LeZergan. Inherited community code and the original game retain their own authorship and licenses. The port has not had a complete playthrough. Expect unexpected bugs. Plants vs. Zombies 2 — PS Vita Unofficial PS Vita loader for the Android 4.5.2 ROW build of Plants vs. Zombies 2. Download — Report a problem — Discord Loader PvZ2 1.1 RC11 (452-v1.1-rc11) Supported Android set 4.5.2 ROW, version 147, ARMv7 Data path ux0:data/pvz2 Licence MIT, with third-party notices No Android game files are included. You must supply your own matching libPVZ2.so and main OBB. Current release status 1.1 RC11 beta is available. Install its VPK over the previous version and keep your existing game files and saves. RC5–RC11 fix audio queue locking, configuration startup races, JNI string handling, and a texture-worker handoff identified in a hardware loading freeze. RC11 also reuses matching shader pairs within a session and moves periodic log writes off the game thread. The latest RC10 hardware log reaches 15,300 frames without a stall report. RC11 hardware performance is not yet measured; busy waves and some transitions still need work. See the RC11 release notes and performance evidence. Requirements Install the following on a homebrew-enabled Vita before the loader: Component Location Notes VitaShell — Used to install the VPK an","listed_at":1789166733,"md_updated_at":1789166733},{"full_name":"Rayekkk/LegionGo2Companion","url":"/r/Rayekkk/LegionGo2Companion","github":"https://github.com/Rayekkk/LegionGo2Companion","title":"LegionGo2Companion","tagline":"Decky plugin for Legion Go 2 on SteamOS, featuring TDP and CPU profiles, battery protection, gyro and touchpad tools, RGB, button remapping, brightness and HDR fixes, and Wi-Fi optimization.","source":"trawl","tier":"found","license":"BSD-3-Clause","stars":5,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Decky plugin for the Legion Go 2 on SteamOS covering TDP profiles, gyro, RGB and button remapping, signed off \"Vibe coded with AI assistance\". It carries the BSD-3-Clause license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Power, battery protection, gyro, haptics, lighting, buttons, screen brightness, HDR and Wi-Fi in one Steam overlay. The Legion Go 2 controls you use every day, with saved profiles and automatic recovery after wake. Features · Requirements · Installation · Usage · How it works · Troubleshooting Features TDP presets and custom limits Five presets, sustained power and burst headroom, with a live package-power reading Game and charging profiles Separate TDP settings for each game and its battery/AC states, applied without opening the menu CPU Boost and EPP Independent CPU controls, saved and checked against the kernel after each change Controller haptics Vibration intensity and modes, touchpad feedback, a test action and per-game vibration profiles Joystick lighting Solid, Breathing, Color cycle and Rainbow effects, with colour, brightness and supported speed controls Power-button light A separate saved toggle on the verified Legion Go 2 firmware Extra-button remapping Independent actions for the left controller's Desktop and Page buttons, including F1-F12 and Disabled Controller gyro Enable gyro reporting through the existing Lenovo driver and choose the left, right or combined controller source Gyro and touchpad diagnostics A short, passive test compares native controller data with the virtual controller output available to Steam Battery protection The firmware's Long Life charging mode, with the previous normal or fast charging mode preserved OLED brightness fix Fixes screen b","listed_at":1789166732,"md_updated_at":1789166732},{"full_name":"BenConAu/Lockjaw","url":"/r/BenConAu/Lockjaw","github":"https://github.com/BenConAu/Lockjaw","title":"Lockjaw","tagline":"Microkernel based OS written in Rust","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Lockjaw, a Rust capability microkernel that boots on a Raspberry Pi 4B; \"100% of the source code in this repository was written by Claude (Anthropic's AI) under direction from Ben.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Lockjaw A capability-based microkernel written in Rust, targeting AArch64 (ARMv8-A). Runs on QEMU virt machine and boots end-to-end on a real Raspberry Pi 4B from the same binary — DTB-driven platform discovery, GICv2 + spin-table SMP, kernel image linked at a fixed higher-half VA with the load physical address discovered at runtime. Inspired by seL4 and Zircon, but with its own object model. Why this exists Lockjaw is part pet project, part experiment, part tutorial. 100% of the source code in this repository was written by Claude (Anthropic's AI) under direction from Ben. Ben writes prompts, plans, design feedback, and review verdicts; Claude writes the code. The project is staged to demonstrate three claims: Claude can write OS-level systems software. Page fault handlers, exception vectors, context-switch assembly, MMU bring-up, IPC state machines, DMA cache-coherence envelopes, the SDHCI driver that reads from a real Pi 4B SD card — the things that have historically demanded years of low-level expertise to get right. The commit log is the proof. An OS can be written in Rust with minimal hand-written asm! and minimal unsafe. The whole kernel binary contains a handful of audited assembly compound transactions (boot trampoline, MMU enable, higher-half pivot, EL0 drop, exception save/restore, context switch) and a handful of audited unsafe helpers (cache maintenance, GIC IAR read, the MMIO atomic-writer transmute, the kernel-image relink machinery). Adding a new audited site ","listed_at":1789166730,"md_updated_at":1789166730},{"full_name":"8lall0/racccoon","url":"/r/8lall0/racccoon","github":"https://github.com/8lall0/racccoon","title":"racccoon","tagline":"A small RISCV-64 toy microkernel written in C3","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"C3","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Racccoon, a small RISC-V microkernel written in C3; its README says \"Almost all of it was written by an LLM agent, at my direction\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Racccoon Racccoon (yes, with three cs — for C3) is a small microkernel for 64-bit RISC-V (rv64imafdc, hardware floating point), written entirely in C3 with no libc. It began as a walk through OS in 1,000 Lines as an experiment in how far C3 can be pushed for kernel and bare-metal work. It has since grown well past 1,000 lines and past the tutorial: a preemptive microkernel with user-space drivers, a Plan 9-style IPC and namespace layer, three filesystems, a USB host stack, and a port to real hardware. This is a vibecoded experiment. Almost all of it was written by an LLM agent, at my direction, as a way to explore architectural choices — a Plan 9-style microkernel, path-based (not fd-based) filesystem IPC, everything-is-a-user-server, board abstraction as the only platform seam, language toolchains (TinyCC, Go) self-hosting on top — and to see how a from-scratch OS holds up under that workflow. It is not audited, not hardened, and not meant for anything real. Treat the code as a sketch of the ideas, not a reference implementation of them. Targets Target Machine Notes racccoon QEMU virt virtio-mmio block / net racccoon-duo Milk-V Duo Sophgo CV1800B, single T-HEAD C906 core; boots as S-mode payload under the stock OpenSBI Sv39 paging, one hart. The second C906 core on the Duo (no MMU) is not a target. What works Scheduler — preemptive, timer-interrupt driven, round-robin. Processes — rfork (RFPROC = process, RFMEM = shared address space = thread), exec (ELF64 or flat binary), e","listed_at":1789166729,"md_updated_at":1789166729},{"full_name":"jherskovic/iSCSI","url":"/r/jherskovic/iSCSI","github":"https://github.com/jherskovic/iSCSI","title":"iSCSI","tagline":"A modern iSCSI initiator for macOS 26/27 on Apple Silicon. FSKit + DiskImages backend; mounts LUNs as APFS with no Terminal.","source":"trawl","tier":"found","license":"Apache-2.0","stars":6,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an iSCSI and NVMe/TCP initiator for macOS on Apple Silicon whose README says \"The vast majority of the code here is written by Claude Code using Opus 5 and Fable 5.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"macOS iSCSI Initiator A modern iSCSI and NVMe/TCP initiator for macOS 26/27 on Apple Silicon. macOS ships no initiator for either; the old open-source iSCSI option is a dead kext. This project puts the iSCSI/TCP and NVMe/TCP protocol engines in a user-space Swift daemon and presents the LUN or namespace as a real block device. The current FSKit/hdiutil backend in this project is a bit hacky, but it works: you can mount LUNs as APFS, with the block device provided by Apple's DiskImages framework. A DriverKit virtual SCSI HBA is the eventual goal once Apple lifts some very real throughput limits (see docs/architecture.md). AI warning The vast majority of the code here is written by Claude Code using Opus 5 and Fable 5. It was closely supervised, but most of the architecture and discovery was by trial-and-error, because lots of things that should work, don't. Most of that is purposefully left in the repo to act as documentation and memory. Claude also wrote most of the documentation, although I reviewed and edited it. Repo/development There are a large number of test/integration scripts that rely on my exact setup. I used two disposable VMs on my dev machine to test ideas and iterate. One had XCode installed, SIP disabled, extension barriers disabled, etc. This was used mostly for initial testing and letting Claude try out ideas harmlessly. The other was as 'virgin' as possible in order to figure out what required user prompts, whether notarization worked, etc. You'll see plenty","listed_at":1789166727,"md_updated_at":1789166727},{"full_name":"jxy-s/hyprv","url":"/r/jxy-s/hyprv","github":"https://github.com/jxy-s/hyprv","title":"hyprv","tagline":"A modern Hyper-V Manager & Remote Desktop client for Windows.","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tabbed Hyper-V manager and Remote Desktop client for Windows whose README says \"Vibe-coded — read this before you trust it\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"hyprv A modern Hyper-V Manager & Remote Desktop client for Windows. Vibe-coded — read this before you trust it hyprv is ~100% AI-generated. It was designed and written end-to-end by Claude through conversational prompting — no human has reviewed the majority of the source code. It works well for the author's daily Hyper-V use, but treat it accordingly. Use at your own risk. (See the MIT license — it ships \"AS IS\".) What it is hyprv is a from-scratch alternative to the built-in Hyper-V Manager, focused on the thing that tool does worst: actually using your VMs. Every virtual machine — and every remote machine — opens as a tab with a real, full-featured RDP session: enhanced-mode graphics, clipboard, dynamic resize, per-monitor DPI scaling. Around that sits a clean, fast, keyboard-friendly UI for managing VM state and hardware. Features Tabbed VM consoles — each VM opens in its own tab with a full RDP session (basic and enhanced mode), clipboard and Ctrl+Alt+Del support, and automatic DPI scaling that matches the host. Remote Desktop hosts — connect to any RDP machine, not just VMs, in the same tabbed UI. Credentials are prompt-every-time; nothing is ever stored or sent by hyprv. Tear-away tabs & multi-window — drag a VM out into its own window, or drop it back onto another. Full VM management — a new-VM wizard, a hardware editor (CPU, memory, disks, network, COM, security/vTPM, boot order…), start / stop / save / checkpoint — all from a compact UI. Light / dark / black","listed_at":1789166725,"md_updated_at":1789166725},{"full_name":"jftuga/mac-screen-search","url":"/r/jftuga/mac-screen-search","github":"https://github.com/jftuga/mac-screen-search","title":"mac-screen-search","tagline":"macOS CLI tool that captures screenshots, performs OCR to find text, and highlights or redacts matches","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a macOS CLI that screenshots, OCRs, and highlights or redacts matching text, whose README says \"This project was vibe-coded with Claude Opus 4.6.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"mac-screen-search A macOS CLI tool that captures a screenshot of the entire screen, performs OCR to find all instances of a search term, draws colored rectangles around each match, and opens the annotated image in Preview. It can also process existing image files in batch using a file glob pattern, making it useful for redacting or blurring sensitive text across many screenshots at once. screenshot mode redact mode park dollar AI Disclaimer This project was vibe-coded with Claude Opus 4.6. While it has been tested on the author's system, it interacts directly with macOS screen capture, image processing, and file I/O, the author makes no guarantees and assumes no responsibility for unintended behavior, missed redactions, or any other issues arising from its use. Use at your own risk and always verify redactions before sharing sensitive screenshots. Requirements macOS (uses ScreenCaptureKit, Vision, and CoreGraphics) Screen Recording permission (for screenshot mode) Swift compiler (only needed if compiling from source; otherwise install via Homebrew or download a release) Installation brew tap jftuga/tap brew install jftuga/tap/mac-screen-search To upgrade to a newer version: brew update brew upgrade mac-screen-search Manual download Download the macOS arm64 binary from the releases page: tar -xJf mac-screen-search_v*.tar.xz cd mac-screen-search_v*/ sudo cp mac-screen-search /usr/local/bin/ Building swiftc -o mac-screen-search mac-screen-search.swift Usage mac-screen-search [-r","listed_at":1789166724,"md_updated_at":1789166724},{"full_name":"rms80/gsworkspace","url":"/r/rms80/gsworkspace","github":"https://github.com/rms80/gsworkspace","title":"gsworkspace","tagline":"Workspace application","source":"trawl","tier":"found","license":"Unlicense","stars":7,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an infinite canvas workspace web app whose README says \"all code files in the project were fully generated by Claude Code\". It carries the Unlicense license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"gsworkspace An infinite canvas webapp for combining text, images, videos, and AI-generated content in a freeform workspace. Can be deployed frontend-only, as a local client/server app, or with a remote server. Design by rms80, Coding by Claude Code Check out gsworkspace.com for more info, and a functional demo of Offline mode. Development Note This project began as an experiment to see how far I could get trying to build a nontrivial web app using Claude Code (CC), without as little human intervention in the code as possible. I have provided architectural guidance and debugging help (ie breakpoints, console investigation, etc), but CC has generated all the code. Literally all of it. This is not a project where I gave CC a high-level spec and asked it to build the entire app. We have gone through hundreds of small incremental steps, adding features and fixing bugs over time. I have guided CC through significant refactors and architectural redesigns. CC and I have collaborated on many planning steps for larger chunks of work. I have accumulated the various TODO.md plans in user_notes/TODO_history.md. I have also kept a lot of my own notes in ryan_notes_to_self.txt. Overview gsworkspace provides an infinite pan/zoom canvas where you can: Drop and arrange text blocks, images, and videos Select / move / resize objects Use AI prompts to generate text, images, or HTML content Built-in non-destructive editing for Images (crop) and Videos (crop/trim/speed/mute) Export and Import entir","listed_at":1789166723,"md_updated_at":1789166723},{"full_name":"SixOfFive/infiniteModel","url":"/r/SixOfFive/infiniteModel","github":"https://github.com/SixOfFive/infiniteModel","title":"infiniteModel","tagline":"Load your LLM models across cpus and gpus - vibe coded science experiment gone wild","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":13,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a homebrew distributed LLM inference engine that splits one model across a fleet of machines, described as a \"vibe coded science experiment gone wild\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"∞ InfiniteModel Run a large language model that's too big for any single machine — by splitting it across the computers you already have. InfiniteModel is a from-scratch distributed LLM inference engine. One controller machine pools the memory and compute of a fleet of worker machines (any mix of GPU and CPU boxes, Windows or Linux) and splits a single transformer model's layers across them over a hand-rolled plain-TCP transport. It's built on Hugging Face transformers + plain PyTorch — no vLLM, TGI, Ray, or torch.distributed — so the same code runs on Windows and Linux. It speaks Ollama-, OpenAI-, and Anthropic-compatible HTTP APIs, so tools you already use (Ollama clients, OpenAI SDKs, Claude Code) can point at the cluster unchanged, plus a live web dashboard. Personal research project — expect rough edges. Hugging Face safetensors is the native format; a model that ships only as a single-file GGUF is converted to safetensors automatically at add time (split multi-part GGUFs aren't supported) — see docs/GGUF.md. A note from the author: I'm fairly new to actually using git, so please go easy on me if the history or workflow isn't textbook — I'm still learning the etiquette. I built this for my own homelab and figured there might be demand for something like it out there, so I'm putting it up in case it's useful to someone else. Issues, suggestions, and patient corrections are all welcome. What it does Fits big models by pooling machines. Pipeline (layer-split) parallelism ov","listed_at":1789166721,"md_updated_at":1789166721},{"full_name":"lvbaoshigao/FFmpeg_plus_plus","url":"/r/lvbaoshigao/FFmpeg_plus_plus","github":"https://github.com/lvbaoshigao/FFmpeg_plus_plus","title":"FFmpeg_plus_plus","tagline":"想在寝室用MP4看喜欢的电影，却一直没有足够方便的工具，这一直是我的遗憾。现在，我为自己(用AI)写了这个工具，它能方便压缩视频画质，让资源顺利在MP4上播放，也算了却这桩心事。","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"Dart","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: FFmpeg++, a cross-platform desktop video, image and audio tool with a node editor, whose README lists \"100% AI-generated via Claude\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🎬 FFmpeg++ 专业视频 / 图片 / 音频处理桌面应用 — 100% AI 生成代码 中文 | English 🤖 本项目 100% 由 AI 生成 📸 软件预览 🎬 主界面 📋 使用演示 中文 📖 概述 FFmpeg++ 是一款基于 Flutter（Material Design 3 前端）+ C++17（共享库后端，通过 FFI 加载）的跨平台桌面视频/图片/音频处理工具。核心功能为蓝图式节点编辑器，支持构建复杂的多步骤处理流程。 支持 Windows、Linux（x64 / ARM64）、macOS（Universal）三大平台。 🏗 架构 ┌──────────────────────────────────┐ │ Flutter 桌面 GUI (Dart) │ ← Material Design 3 │ Dart FFI ↔ libffmpegpp │ ├──────────────────────────────────┤ │ C++17 后端 (dll/so/dylib) │ ← 共享库模式（FFI 轮询） │ subprocess → ffmpeg / ffprobe │ ├──────────────────────────────────┤ │ FFmpeg / FFprobe │ ← 外部依赖（用户自行安装） └──────────────────────────────────┘ ✨ 功能 模块 说明 🎬 项目 多视频导入、ffprobe 自动探测、缩略图预览 📋 处理队列 顺序批量处理、实时进度解析 🧩 节点编辑器 蓝图式 DAG 画布，25+ 节点类型，构建复杂多步骤处理流程 🎞 视频转码 17+ 编码器（H.264/H.265/AV1/VP9/SVT-AV1），GPU 加速（NVIDIA/AMD/Intel） 🎵 音频处理 转码 / 变速 / 音量调整 / 动态压缩 / 元信息编辑 / 提取音频（带预览播放） 📝 字幕 烧录外挂 SRT/ASS/SSA，拾色器，系统字体选择器（含预览） 📷 帧提取 单帧 / 范围分帧 / 全部分帧 ✂️ 片段截取 时间范围截取，级联时长约束 🖼 图片处理 格式转换 / 裁剪 / 旋转 / 缩放 / 亮度 / 噪点 / 锐化 / 降噪 / 通道提取 🎬 视频裁剪 交互式选区工具，支持多选区、拖拽调整、保留/移除模式 🔗 合并媒体 多文件顺序合并，图片序列合成视频 🧠 命令 手动输入 ffmpeg 命令 + 快捷模板 + 参数参考 🤖 AI 助手 内置 AI 聊天面板，自然语言描述需求自动配置节点 ⚙️ 设置 暗/亮主题、字体、主题色、背景图片、编辑模式切换 🧩 节点编辑器 节点编辑器是 FFmpeg++ 的核心。详见 NODE_EDITOR.md。 无限画布，支持平移缩放 拖拽节点，自由连线 右键添加节点 / 删除连线 25+ 节点类型覆盖视频、音频、图片处理 自动验证（环路检测、类型冲突、时长约束） 智能合并：音视频处理 + 字幕 = 单条 ffmpeg 命令 逻辑块：循环处理支持 调试覆盖层显示执行计划 多源文件节点 = 多个独立任务 📦 安装 平台 下载 Windows Releases → FFmpeg++_v*_setup.exe Linux x64 Releases → ffmpegpp_*_amd64.deb Linux ARM64 Releases → ffmpegpp_*_arm64.deb macOS Release","listed_at":1789166719,"md_updated_at":1789166719},{"full_name":"jkeifer/0x00c0dec5","url":"/r/jkeifer/0x00c0dec5","github":"https://github.com/jkeifer/0x00c0dec5","title":"0x00c0dec5","tagline":"A static webapp to explore data file format design","source":"trawl","tier":"found","license":"MIT","stars":7,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an interactive web app for learning how data file formats are built, whose README says \"This tool is vibecoded. YMMV.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"0x00C0DEC5 An interactive web tool for learning how data file formats are constructed. Every file format -- Parquet, GeoTIFF, Zarr, HDF5 -- solves the same fundamental problems: how to structure data, encode it efficiently, and provide enough metadata to read it back. This tool lets you discover those problems firsthand by building a format from scratch. NoteThis tool is vibecoded. YMMV. How it works Define a dataset, then watch it transform step-by-step from human-readable values into raw bytes on disk. The transformation pipeline covers: Chunking -- splitting data into spatial slices Interleaving -- row-oriented (BIP) vs. column-oriented (BSQ) byte layout Codecs -- delta encoding, byte shuffle, bit rounding, scale/offset, RLE, LZ compression Metadata assembly -- JSON or binary serialization of structural and user-defined metadata File assembly -- magic numbers, metadata placement, chunk ordering, partitioning A dual-pane comparison view lets you inspect any two pipeline stages side-by-side, with cross-pane hover tracing that links human-readable values to their encoded bytes. Use cases Live talks: walk an audience through building a file format interactively, letting them choose options (row vs. column orientation, which codecs to apply) and see the consequences in real time. Self-guided exploration: experiment independently with different format designs to understand why real-world formats make the choices they do. Development Requires Node.js. npm install npm run dev # st","listed_at":1789166717,"md_updated_at":1789166717},{"full_name":"tosiabunio/pyelectroman","url":"/r/tosiabunio/pyelectroman","github":"https://github.com/tosiabunio/pyelectroman","title":"pyelectroman","tagline":"PyElectroMan is a Python port of 1992 PC game Electro Man (aka Electro Body, a cult game in Poland where it was developed).  ","source":"trawl","tier":"found","license":"MIT","stars":8,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Python port of the 1992 DOS platformer Electro Man whose README says \"Nearly all recent commits have been written by Claude Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"PyElectroMan A Python port of Electro Man (1992), also known as Electro Body — a cult classic platformer developed in Poland by xLand. About PyElectroMan is an ongoing effort to recreate the original DOS game in Python. The project originated as a personal exercise to improve Python skills rather than as a path toward a commercial release. The port is based on the original (unreleased) source code and uses art assets from the game, licensed under CC BY-SA. The original game can be played online at: DOSGames.com RetroGames.cz The DOSBox version, ready to run on modern Windows, is available for free at GOG.com. Project History Development began years ago but went dormant for roughly 15 years due to the complexity of the original codebase — 30+ years old, difficult to parse, and challenging to translate accurately. With the advent of AI-assisted development, the project has been revived as a testbed for Claude Code. Nearly all recent commits have been written by Claude Code. Development Philosophy Conversion accuracy over code quality — the goal is to faithfully recreate original behavior, even when the source code is convoluted Incremental progress — given the complexity, exact 1:1 conversion may not be achievable AI-assisted development — exploring what's possible when AI handles the heavy lifting of legacy code translation Status 🚧 Work in progress — not intended as a playable release at this time. Useful debug info PyElectroMan Debug Shortcuts License Code: MIT License Art ","listed_at":1789166715,"md_updated_at":1789166715},{"full_name":"sjvrensburg/railreader2","url":"/r/sjvrensburg/railreader2","github":"https://github.com/sjvrensburg/railreader2","title":"railreader2","tagline":"Desktop PDF viewer optimised for high magnification viewing.","source":"trawl","tier":"found","license":"MIT","stars":10,"language":"C#","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a desktop PDF viewer for sustained high-magnification reading whose README says \"RailReader2 was built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"railreader2 Desktop PDF viewer optimised for high magnification viewing with AI-guided \"rail reading\". Built with .NET/Avalonia, PDFtoImage (PDFium) for PDF rasterisation, SkiaSharp for GPU-accelerated rendering, and Docling Heron-INT8 (ONNX) for layout detection. Download · Microsoft Store · Website · User Guide License RailReader2 is licensed under the MIT License. Versions prior to 3.0.0 were released under the GNU General Public License v3 (GPLv3). Starting with version 3.0.0, the project is licensed under the MIT License. As the sole author and copyright holder, this relicensing applies to all new releases going forward. Previous releases remain available under the terms of the GPLv3. Why I built this As a visually impaired user, I need a PDF viewer that works comfortably at high magnification for sustained reading. The tech industry rarely builds for this \"missing middle\": the market is too niche for standard software companies, and tools designed for full blindness aren't appropriate when you have poor-but-usable vision. High magnification introduces real navigation challenges: context loss, inefficient scrolling, and UI elements that break at zoom. RailReader2 addresses these by using AI layout analysis to guide line-by-line reading through detected text blocks, like a typewriter carriage return across the page. RailReader2 was built with Claude Code. It is an example of how AI-assisted development can enable niche accessibility tools that would otherwise never exist.","listed_at":1789166713,"md_updated_at":1789166713},{"full_name":"jnsgruk/booklog","url":"/r/jnsgruk/booklog","github":"https://github.com/jnsgruk/booklog","title":"booklog","tagline":"B{ook}log is a self-hosted, multi-user book tracking platform for avid readers. Each user maintains a personal library and wishlist, with per-user readings, stats, and timeline.","source":"trawl","tier":"found","license":"Apache-2.0","stars":10,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hosted multi-user book tracking app whose README says \"This application was built almost entirely with Claude Code.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"B{ook}log is a self-hosted, multi-user book tracking platform for avid readers. Each user maintains a personal library and wishlist, with per-user readings, stats, and timeline. Booklog features an LLM-powered extraction feature, which enables it to automatically fill book and author information using a photo of a book cover. Booklog ships as a single Rust binary that serves a web UI, a REST API, and a CLI client. The application uses SQLite as a backend, and will automatically create and migrate the database on start-up. NoteThis application was built almost entirely with Claude Code. I used this project as an excuse to explore the current state of the art in agentic coding tools. I've reviewed much of the code, and I'm largely responsible for the rules and the layout of the repository but nonetheless the vast majority of the code was written by Claude and Opus 4.6. Quick Start (Demo) Optionally, sign up for OpenRouter to enable AI-powered book cover extraction. Then create a docker.env file: # Optional: API key from OpenRouter for AI extraction BOOKLOG_OPENROUTER_API_KEY=sk-or-... # I've had good results with Gemini models, but you can try 'openrouter/free' to experiment BOOKLOG_OPENROUTER_MODEL=google/gemini-3-flash-preview See the full list of configuration options below. Once the .env file is ready, start the container using the environment file: # Create a data directory to store the database mkdir data # Run the container docker run \\ --rm \\ -p 3000 \\ --env-file docker","listed_at":1789166712,"md_updated_at":1789166712},{"full_name":"HelgeSverre/token","url":"/r/HelgeSverre/token","github":"https://github.com/HelgeSverre/token","title":"token","tagline":"Minimal text editor in Rust using the Elm Architecture","source":"trawl","tier":"found","license":"MIT","stars":9,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1000,"vulnerable":13},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a multi-cursor code editor inspired by JetBrains IDEs, \"Vibe-coded in Rust, using Amp Code.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Token - A Multi-Cursor Text Editor Multi-cursor, code editor inspired by JetBrains IDEs, Vibe-coded in Rust, using Amp Code. The repository includes a reviewed archive of 171 Amp conversations that informed the project. Installation Pre-built Binaries Download the latest release for your platform from token-editor.com or GitHub Releases. Building from Source Requires the Rust toolchain. git clone https://github.com/HelgeSverre/token cd token just setup # Install dependencies just release # Build optimized binary just run # Run the editor For development: just dev # Run debug build (faster compile) just watch # Start bacon watch mode Windows MSI packaging requires Windows, the Rust MSVC toolchain, PowerShell 7, and WiX Toolset 3.14 (candle.exe/light.exe on PATH, or the installation directory in WIX). Run just bundle-windows; the installer is written to target/x86_64-pc-windows-msvc/release/bundle/msi/Token.msi. The CI-only smoke script installs and uninstalls the package, so run it only on a disposable Windows system. Quick Start Open a file: token path/to/file.rs or use Cmd+O Open a folder: token path/to/project/ or use Cmd+Shift+O Command palette: Cmd+Shift+A for all available commands Toggle sidebar: Cmd+1 to show/hide the file tree Command line token returns to the shell right away and opens files in the editor that is already running; a folder always gets its own window. Invocation Effect token src/main.rs:42:7 Open at line 42, column 7 (--line/--column also work) token -","listed_at":1789166426,"md_updated_at":1789166426},{"full_name":"Olsro/jellydinosaur","url":"/r/Olsro/jellydinosaur","github":"https://github.com/Olsro/jellydinosaur","title":"jellydinosaur","tagline":"A very compatible, fast, no bullshit, and less than 100kb JellyFin front-end that you can self host","source":"trawl","tier":"found","license":"MIT","stars":15,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A tiny ES3 Jellyfin web client for old browsers, with a Flash player fallback, whose README credits \"AI-assisted code\" with \"a lot of Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"JellyDinosaur A no bullshit, very compatible, super fast, and JavaScript ES3 compliant Jellyfin client (front-end). It is a static < 100kb HTML webpage, which can be uploaded to any HTTP web server or hosted locally. Uploading hls.min.js is necessary only to stream HLS from most modern browsers which lacks native support. The 8kb jellydinosaur-flash-player.swf is a minimalistic AS3 streaming flash client and is useful if your browser has a Flash plug-in. To host your instance of JellyDinosaur locally and quick, use on the upload folder: python3 -m http.server 20005 You may want to setup this app with an integrated reverse proxy on your LAN which you can do with: node serve-with-proxy.js Reverse proxy is especially useful if you need very old operating systems/browsers which lacks modern TLS to access any kind of Jellyfin server requiring SSL. Screenshots Try it right now You can try it from the GitHub page from here if your browser supports modern HTTPS: https://olsro.github.io/jellydinosaur/upload Credentials: Jellyfin instance: https://demo.jellyfin.org/unstable Login : demo Password : (don't put anything) Tech stack Pure HTML/ES3 JS (index.html) (Optional) A flash AS3 app for streaming. I already compiled this program as an swf in the repo so you don't have to mess with the Flex SDK if you don't intend to modify it. (Optional) NodeJS if you want to use the easy SSL escalation reverse proxy script (serve-with-proxy.js) and the build-minify.js minification script I provid","listed_at":1789166419,"md_updated_at":1789166419},{"full_name":"KSXGitHub/perfectionist","url":"/r/KSXGitHub/perfectionist","github":"https://github.com/KSXGitHub/perfectionist","title":"perfectionist","tagline":"Additional linting rules for Rust projects","source":"trawl","tier":"found","license":"MIT","stars":13,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A set of extra Dylint lint rules for Rust projects; its FAQ admits \"This code base is almost entirely AI-generated. Or in other words, vibe-coded.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"perfectionist Additional linting rules for Rust projects. Rules See the homepage or the rules/ directory for all the rules implemented so far. See planned-rules/ for all the rules not yet implemented. Usage Install Dylint, then add perfectionist to your workspace's dylint.toml: [workspace.metadata.dylint] libraries = [ { git = \"https://github.com/KSXGitHub/perfectionist\" }, ] Run the lints: cargo dylint --all -- --all-targets Controlling rules Each lint registers under the perfectionist tool namespace. See CONTROLLING_RULES.md for how to change a rule's level per call site or project-wide, and how to enable or disable rules globally. Configuration dylint.toml accepts a crate-wide [perfectionist] table and per-rule [\"perfectionist::<rule>\"] tables. See CONFIGURATION.md for the full schema. Development Prerequisites: Rustup Just cargo-dylint and dylint-link are not part of the cargo dependency graph but their ABI is coupled to it. They're installed into a workspace-local .dev-tools/ directory (pinned to the dylint_linting version in Cargo.lock) by: just install-dev-tools Run this once after cloning, and again whenever dylint_linting is bumped. Every other just recipe prepends .dev-tools/bin to PATH, so subsequent commands use the pinned binaries automatically. Run the following command to check everything: just all Frequently Asked Questions Why does this code base look so ugly? This code base is almost entirely AI-generated. Or in other words, vibe-coded. Why vibe-code? ","listed_at":1789166415,"md_updated_at":1789166415},{"full_name":"ZythDr/Wuddle","url":"/r/ZythDr/Wuddle","github":"https://github.com/ZythDr/Wuddle","title":"Wuddle","tagline":"World of Warcraft Launcher and Manager (Addons, DLLs, MPQs) primarily for Vanilla 1.x, TBC 2.x, and WotLK 3.x clients.  Wuddle is entirely vibe-coded but I try my best to make sure it doesn't suck.   Expect a few bugs and rough edges as it's my first attempt at making an app.","source":"trawl","tier":"found","license":"MIT","stars":14,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A desktop launcher and addon manager for legacy World of Warcraft clients; \"Wuddle is entirely vibe-coded but I try my best to make sure it doesn't suck.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Wuddle - Legacy WoW Launcher & Manager A native desktop launcher and manager for legacy World of Warcraft clients. Wuddle brings addon management, DLL mods, MPQ patches, profiles, game launching, updates, and client-specific tools into one application for Vanilla 1.12.1, TBC 2.4.3, and WotLK 3.3.5 clients. Download Latest Stable Release: Linux Windows Antivirus false-positives Warning TL;DR: To avoid issues, every game installation directory where you intend to use the Mods tab should be whitelisted in your antivirus software. If any mod fails to install, or if Windows Defender displays a notification as you install a mod, you need to whitelist your game install folder. Various legacy DLL mods, including projects such as SuperWoW, UnitXP_SP3, and Nampower, are known to trigger false positives in antivirus software. Wuddle warns you before installing known false-positive-triggering mods from the Quick Add list. But since Wuddle performs the download and installation of these mods, your antivirus software may attribute the detection of a threat directly to Wuddle.exe itself. The most reliable solution is to whitelist your game installation folder in your antivirus software. If this sounds sketchy, Wuddle is open source and the code is available here on GitHub for you to inspect. So to simplify the user-experience, Wuddle creates a hidden .wuddle cache/staging folder inside the game installation's root directory. This is where all Mods, Addons, MPQs and so on land when they ","listed_at":1789166123,"md_updated_at":1789166123},{"full_name":"jnsgruk/brewlog","url":"/r/jnsgruk/brewlog","github":"https://github.com/jnsgruk/brewlog","title":"brewlog","tagline":"B{rew}log is a self-hosted coffee logging platform for tracking your roasters, roasts, brews, cafes and brewing gear.","source":"trawl","tier":"found","license":"Apache-2.0","stars":15,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A self-hosted coffee logging platform for roasters, roasts, brews and cafes; the owner notes \"This application was built almost entirely with Claude Code.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"B{rew}log is a self-hosted specialty coffee logging platform optimised for filter brewing enthusiasts. B{rew}log can be used for tracking roasters, roasts, brews, cafes and brewing gear. B{rew}log features an LLM-powered \"Bag Scanning\" feature, which enables it to automatically fill roaster and coffee information using a photo of a bag. It also supports \"check-ins\" to log coffee enjoyed in a cafe. B{rew}log ships as a single Rust binary that serves a web UI, a REST API, and a CLI client. The application uses SQLite as a backend, and will automatically create and migrate the database on start-up. NoteThis application was built almost entirely with Claude Code. I used this project as an excuse to explore the current state of the art in agentic coding tools. I've reviewed much of the code, and I'm largely responsible for the rules and the layout of the repository but nonetheless the vast majority of the code was written by Claude and Opus 4.6. Quick Start (Demo) Before you start, you'll need to sign up for Openrouter and Foursquare Places and get API keys for both. Then create a docker.env file: # You'll need an API key from OpenRouter BREWLOG_OPENROUTER_API_KEY=sk-or-... # I've had good results with Gemini models, but you can try 'openrouter/free' to experiment BREWLOG_OPENROUTER_MODEL=google/gemini-3-flash-preview # FourSquare Places API key for location searching BREWLOG_FOURSQUARE_API_KEY=fsq3... You can see the full list of configuration options below. Once your .env file i","listed_at":1789166120,"md_updated_at":1789166120},{"full_name":"dandersondev/visual-notes","url":"/r/dandersondev/visual-notes","github":"https://github.com/dandersondev/visual-notes","title":"visual-notes","tagline":"A visual workspace built on Obsidian's Canvas format: nestable freeform boards, icon tile grids, kanban boards, sticky notes, checklists, columns, drawing with pen, labels, reactions, and more..","source":"trawl","tier":"found","license":"MIT","stars":22,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["plugin"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: An Obsidian plugin for Canvas boards with kanban, sticky notes and drawing, whose owner says \"Most of the code is written by Claude; I direct it, test it, and decide what ships.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Visual Notes Visual notes, planning and organisation. Boards that live in your Obsidian vault as real Canvas files — so they open in Obsidian's own Canvas view, nest inside each other, and outlast the plugin that made them. A place to think in space rather than in a list, the way you would in Milanote, Notion or Trello: plan a project across a wall of sticky notes, keep a reading list as icon tiles, run a kanban board, draw over the top of it, and connect the pieces with labelled lines. How is this different from Obsidian's Canvas? It is Canvas — the same spec-compliant .canvas file, sitting in your vault, openable in the native view. Visual Notes adds what Canvas leaves out: live collaboration, storyboards, kanban boards, tables, checklists, calendars, columns, group frames, pen and highlighter drawing, and boards that nest inside one another. Experimental collaboration Work on the same board at the same time. Live cursors, live selections, live edits — everyone watching the same canvas change as it changes. No cloud, no account, and no server but yours. One desktop participant hosts the room from inside Obsidian and everyone else joins with an invitation, over a private network you already control. Nothing is routed through Visual Notes, because there is nothing to route it through. Everyone keeps an ordinary .canvas file in their own vault, so a board is still yours when the session ends and the host goes offline. Cards, connections, drawings, nested boards, kanban boards,","listed_at":1789166116,"md_updated_at":1789166116},{"full_name":"joonlab/MacPilot","url":"/r/joonlab/MacPilot","github":"https://github.com/joonlab/MacPilot","title":"MacPilot","tagline":"Turn your Mac into a wireless trackpad, keyboard & Stream-Deck — controlled from any phone's browser. No app install. Open source, built with Claude Code.","source":"trawl","tier":"found","license":"MIT","stars":23,"language":"C#","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A Mac menu-bar helper that turns any phone browser into a wireless trackpad, keyboard and Stream-Deck-style panel, \"Built end-to-end with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"MacPilot Turn your Mac into a wireless trackpad, keyboard & Stream-Deck — controlled from any phone's browser. No app to install. Open source. Built end-to-end with Claude Code. 아이폰·아이패드·안드로이드 어디서든 브라우저만 열면 맥을 마우스/키보드/단축키 패널로 조작. 앱 설치 0. 🎬 Demo ▶︎ Open as mp4 file What is this? MacPilot runs a tiny menu-bar helper on your Mac that serves a web app over your local Wi-Fi. Open the URL on any phone or tablet and you get a full trackpad + keyboard + customizable shortcut/macro deck — all in the browser. No App Store. No install on the phone. Just open a URL (or scan the QR in the menu bar). Works on iOS, iPadOS, Android — anything with a browser. Free & open source — a replacement for paid apps like Remote Mouse, Unified Remote, etc. 원래 비슷한 기능은 유료 앱을 결제해서 써야 했지만, 이건 무료 오픈소스이고 폰에 앱을 깔 필요조차 없습니다. ✨ Highlights 🖱️ Real-trackpad feel — cursor with pointer acceleration, tap / double-tap / right-click, tap-and-drag, two-finger scroll with momentum, pinch-zoom, and three-finger swipes (Mission Control / spaces-style) ⌨️ Full keyboard — types real text incl. Korean & emoji (Unicode), plus held-modifier combos (hold ⌘ and type) 🎛️ Stream-Deck-style deck — folders, swipeable pages, and buttons for: Shortcuts up to 4-key combos (⌘⌃⇧⌥ + any key) Text snippets · App / link / deep-link launch (pick from your installed apps, with icons) Multi-step macros (key → delay → text → launch, in sequence) Per-button icons & colors, drag-to-reorder, move between folders → covers much of","listed_at":1789166113,"md_updated_at":1789166113},{"full_name":"git4sim/HA-Toniebox","url":"/r/git4sim/HA-Toniebox","github":"https://github.com/git4sim/HA-Toniebox","title":"HA-Toniebox","tagline":"Unofficial Home Assistant integration for Toniebox / Tonie Cloud","source":"trawl","tier":"found","license":"MIT","stars":24,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["iot"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: An unofficial Home Assistant integration for the Toniebox kids' audio player, whose README says \"This integration was vibecoded\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"HA-Toniebox Unofficial Home Assistant Integration for Toniebox / Tonie Cloud Inoffizielle Home Assistant Integration für Toniebox / Tonie Cloud WarningDisclaimer — Please read before use / Bitte vor der Nutzung lesen This project has no affiliation with Boxine GmbH (manufacturer of Toniebox / tonies.de) and is neither supported nor endorsed by them. It uses the official Tonie Cloud REST API, which may change at any time without notice. Use at your own risk. No warranty, no support from Boxine. Dieses Projekt steht in keiner Verbindung zu Boxine GmbH und wird von diesen weder unterstützt noch empfohlen. Es nutzt die offizielle Tonie Cloud REST API, die sich jederzeit ohne Ankündigung ändern kann. Nutzung auf eigene Gefahr. 🤖 This integration was vibecoded — developed with AI assistance (Claude by Anthropic) and tested against a real Toniebox account. 🇬🇧 English Features 🧸 Each Creative Tonie as its own device with media player, cover image and chapter list 📻 Each Toniebox as its own device — shows the currently placed figure with name and cover ▶️ Full playback control on the Toniebox 2 (TNG) — play/pause, next/previous chapter, jump to any chapter (browse), volume, and put to sleep — live via the Tonie Cloud's real-time channel 🎚️ Live \"now playing\" — current chapter title, progress bar and play/pause state ⏲️ Sleep timer — start a timer via the toniebox.sleeptimer action, with a sleeptimer_active binary sensor 📊 Sensors for chapter count, total duration, firmware vers","listed_at":1789165820,"md_updated_at":1789165820},{"full_name":"pcoronaf/PhoinixDR","url":"/r/pcoronaf/PhoinixDR","github":"https://github.com/pcoronaf/PhoinixDR","title":"PhoinixDR","tagline":"Open-source data recovery for everyone.","source":"trawl","tier":"found","license":"Apache-2.0","stars":32,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an open-source data recovery engine, CLI and desktop app for NTFS, FAT, exFAT and ext drives, whose docs declare \"Yes, PHOINIX is vibecoded\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Open Source Data Recovery · by @pcoronaf Lost data lives again. Recover lost files. Understand your chances. Download for Windows · Other platforms · Website · Documentation · User guide · Changelog · Documentación en español PhoinixDR (PHOINIX Data Recovery) is a data-recovery engine, command-line tool and desktop application that reconstructs lost data from filesystems, raw media, disk images and damaged storage structures while explaining how likely each recovered object is to be intact. Select source → Scan → Find lost data → Assess recoverability → Preview → Recover Status: early engineering preview. The repository implements milestones M0–M11 of the technical specification: the read-only block layer, MBR/GPT discovery, native NTFS, FAT12/16/32, exFAT and ext2/3/4 readers with undelete (journal-assisted on ext3/ext4), deep scan (signature carving of unallocated space), lost-partition recovery (virtual mounts, no table writes), forensic and virtual-disk image containers (E01, split E01, split RAW, VHD, VHDX, VMDK) with hash verification and exportable recovery reports, evidence-based recovery health, a verified recovery writer, the phoinix CLI and a desktop application (Tauri 2 + React) with sessions, previews and recovery. About the project Where PHOINIX Came From: why a complete, modern recovery platform rather than another utility or a front-end for TestDisk, and why no code is copied from existing projects. Development Declaration: how AI-assisted development is used,","listed_at":1789165816,"md_updated_at":1789165816},{"full_name":"tulustul/bevy-react","url":"/r/tulustul/bevy-react","github":"https://github.com/tulustul/bevy-react","title":"bevy-react","tagline":"A library for rendering bevy_ui elements inside Bevy app using React","source":"trawl","tier":"found","license":"Apache-2.0","stars":25,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A Rust library for building Bevy game UIs with React components, which its owner calls a \"quick, vibecoded proof of concept\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"bevy-react Build bevy_ui interfaces with React. You write components in React/TSX and they render to native Bevy UI through a React Native-style bridge - no web view, no DOM. The JS side stays purely declarative; Rust and Bevy do the heavy lifting. State and interactions flow both ways between your Bevy app and React, and edits hot-reload live while keeping component state. You can play with a live demo here: https://tulustul.github.io/bevy-react/ import { mount } from \"bevy-react\"; import { useState } from \"react\"; function App() { const [n, setN] = useState(0); return ( <node style={{ padding: 20, gap: 12, flexDirection: \"column\" }}> <text>{`Count: ${n}`}</text> <button onClick={() => setN((c) => c + 1)} style={{ backgroundColor: \"#7aa2f7\" }} > <text>+</text> </button> </node> ); } mount(<App />); That's a real component - <node> and <button> render to actual bevy_ui nodes, useState works as you'd expect, and saving the file updates the running app without losing the count. Why bevy-react React, not a bespoke UI DSL. Hooks, components, conditional rendering, lists - everything you already know. Native Bevy UI. No web view, no DOM. Your UI is bevy_ui entities in the same world as your game. Hot reload that keeps state. Edit a component and it re-renders live with hook state and running animations intact. Typed, two-way messaging. React and the ECS talk over typed channels generated straight from your Rus","listed_at":1789165812,"md_updated_at":1789165812},{"full_name":"clarkarch/tfm-tui","url":"/r/clarkarch/tfm-tui","github":"https://github.com/clarkarch/tfm-tui","title":"tfm-tui","tagline":"Modern mouse-first terminal file manager","source":"trawl","tier":"found","license":"MIT","stars":85,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a mouse-first terminal file manager with drag and drop, tabs and image thumbnails, labeled \"Experimental vibecoded software, expect rough edges.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"tfm (terminal file manager) A modern, mouse-first file manager with places sidebar, grid view, drag & drop, image thumbnails and more, right inside your terminal. WarningExperimental vibecoded software, expect rough edges. Don't test on files you can't afford to lose. If you want to test safely, use Podman: podman run --rm -it archlinux bash ImportantThis is still a terminal UI running inside your terminal, expect some visual/behavioral anomalies. Features Click, rubber-band select, right-click menus, inline rename. Drag files between folders (ctrl+drag), out to other apps, or in from outside. Cross-app drag is kitty-only. Places sidebar, GTK bookmarks, recent files, XDG trash with restore, clipboard. Network locations: connect to sftp://, smb://, WebDAV, FTP… shares through gvfs (sidebar or right-click → Connect to Server…); saved connections live in your GTK bookmarks. Credentials are prompted in-app (passwords masked); ssh keys/agent work without prompting. Embedded terminal (right-click → Open Terminal Here). Auto-hide panes: sidebar/preview/terminal collapse to the edge and slide back (animated) when the mouse nears them. Extract and compress archives (right-click). Image/video thumbnails, text syntax highlighting, folder sizes. Type-to-search, tabs, undo/redo, 30+ themes. Requirements Linux. Image thumbnails need the kitty graphics protocol (kitty, ghostty, WezTerm, Konsole…). Without it you get Nerd Font glyphs. Cross-app drag needs kitty. Optional tools (the insta","listed_at":1789165535,"md_updated_at":1789165535},{"full_name":"atxoxx/GameIndex","url":"/r/atxoxx/GameIndex","github":"https://github.com/atxoxx/GameIndex","title":"GameIndex","tagline":"Unify your Steam, GOG, Epic, and DRM-free libraries into a single, fast, native experience — with discovery, sync, activity tracking, and a controller-first Big Picture mode (Beta).","source":"trawl","tier":"found","license":"MIT","stars":36,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["game","library"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1000,"vulnerable":9},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a cross-store PC game launcher that unifies Steam, GOG, Epic and DRM-free libraries with a controller-first mode, a \"Personal project, vibe-coded\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"GameIndex A unified, cross-store game launcher and library manager. Unify your Steam, GOG, Epic, Rockstar, Ubisoft, and DRM-free libraries into a single, fast, native experience — with discovery, deals & news, activity tracking, achievements, Linux/Steam Deck Proton support, a social layer, and a controller-first 10-foot UI. 🎬 Demo Video A quick guided tour of GameIndex — the launcher, unified library, game pages, and more: ▶️ Watch the presentation on YouTube 📑 Table of Contents ✨ Features 🐧 Linux, Wine & Proton 📸 Screenshots 🎬 Demo Video 💡 Inspiration 🛠️ Tech Stack 🚀 Getting Started 📁 Project Structure 🗺️ Roadmap 📌 Status 🤝 Contributing ⚖️ Disclaimer 📄 License 🙏 Acknowledgments ✨ Features Feature Description Unified Library Steam, GOG Galaxy, Epic Games Store, Rockstar, Ubisoft Connect, Humble Bundle, and manual imports in one cohesive grid. Rich Game Pages Hero, metadata, reviews, achievements, screenshots, videos, web links, HowLongToBeat stats, Crackwatch, ProtonDB, Steam community features, and live player counts. Game Notes Per-game Markdown workspace with multiple notes, 7 templates (walkthrough, checklist, boss strategy, build & loadout, quest log, progress journal), search & tag filters, pinning, live preview (write/split), word count, copy/export, and debounced auto-save. Game Versions Detects installed versions from GOG/Epic/Steam manifests or PE metadata and flags newer releases on game pages and in the download modal. Achievements H","listed_at":1789165532,"md_updated_at":1789165532},{"full_name":"pelgraine/Meck","url":"/r/pelgraine/Meck","github":"https://github.com/pelgraine/Meck","title":"Meck","tagline":"Meshcore + Fork = Meck. A fork created specifically to focus on enabling BLE & WiFi companion firmware for the LilyGo T-Deck Pro, T-Deck Max and T5 E-Paper S3 Pro. Created wholly with Claude AI using Meshcore v1.11 code. 100% vibecoded. 100% free.","source":"trawl","tier":"found","license":"MIT","stars":32,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: MeshCore companion firmware for LilyGo T-Deck and e-paper devices, \"Created wholly with Claude AI using Meshcore v1.11 code. 100% vibecoded.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Meshcore + Fork = Meck A fork created specifically to focus on enabling BLE & WiFi companion firmware for the LilyGo T-Deck Pro, LilyGo T-Deck Max & LilyGo T5 E-Paper S3 Pro. Created wholly with Claude AI using Meshcore v1.11 code. 100% vibecoded. Check out the Meck discussion channel on the MeshCore Discord Meck Watch -- the LilyGo T-Watch S3 and T-Watch S3 Plus firmware -- is a separate, private repository, available to monthly GitHub sponsors from $2 USD per month. Contents Supported Devices SD Card Requirements Flashing Firmware First-Time Flash (Merged Firmware) Upgrading Firmware Launcher OTA Firmware Update Path Hash Mode (v0.9.9+) Region Scope (v1.7+) T-Deck Pro Build Variants Keyboard Controls Navigation (Home Screen) Bluetooth (BLE) WiFi Companion Clock & Timezone Channel Message Screen Channel Picker Contacts Screen Sending a Direct Message Roomservers Repeater Admin Screen Trace Route Screen (v1.9+) Rx Log Delete Message History (v1.10+) Per-Channel Notification Preferences (v1.10+) Custom Notification Tones (v1.10+) Games (v1.10+) Private Channels (v1.11+) Channel Sharing via DM (v1.11+) Config Export/Import (v1.11+) Settings Screen Font Styles Compose Mode Symbol Entry (Sym Key) Emoji Picker SMS & Phone App (4G only) Web Browser & IRC Alarm Clock (Audio only) Voice Notes Over LoRa (Audio only) Lock Screen (T-Deck Pro) Shutdown (T-Deck Pro) T-Deck Max Build Variants 4G and Audio at the Same Time Antenna (Internal / External) Buzzer (Vibrate) N","listed_at":1789165514,"md_updated_at":1789165514},{"full_name":"deutsia/deutsia-radio","url":"/r/deutsia/deutsia-radio","github":"https://github.com/deutsia/deutsia-radio","title":"deutsia-radio","tagline":"A privacy-focused, censorship-resistant multinet Android radio player built with Claude Code. Supports I2P, clearnet and Tor streaming.","source":"trawl","tier":"found","license":"Apache-2.0","stars":74,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["mobile","security"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a privacy-focused Android internet radio player that streams over clearnet, Tor and I2P, described by its owner as \"built with Claude Code\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"deutsia radio Privacy-focused, censorship-resistant multinet radio player with Material You. ar | de | es | fa | fr | hi | it | ja | ko | my | pt | ru | tr | uk | vi | zh Updates New Release: v1.8.0 - Android Auto support, bug fixes Features Multinet Support - Clearnet, I2P (.i2p), and Tor (.onion) with custom Darknet API providing Tor/I2P stations Force Tor Modes - Proxy-or-fail enforcement with instant leak prevention 50,000+ Stations - Access RadioBrowser API or add custom streams Material You - Dynamic theming along with 6 color schemes and Dark/Light mode Stream Recording - Record across multiple stations App Lock - Secure your listening with biometric/PIN protection and optional database encrpytion 17 Languages - Multilingual interface support Add any station Support for Tor, Clearnet, I2P and per-station level custom proxy stations Full DNS Protection - No DNS leaks to ISP/System DNS in Force Tor/Proxy modes, backed by 47 unit and integration tests Disable API connections - Completely disable API connections + external cover art and utilize the app with only your local library Privacy & Security Optional Database Encryption via SQLCipher: All local data encrypted at rest (stations, settings, history) App lock with biometric/PIN protection Secure storage for favorites and playback data Credential Protection via Android Jetpack Security: Proxy passwords encrypted with AES-256-GCM Hardware-backed encryption keys (when available) Secure credential storage for custom pr","listed_at":1789165248,"md_updated_at":1789165248},{"full_name":"ncosgray/cuppa_mobile","url":"/r/ncosgray/cuppa_mobile","github":"https://github.com/ncosgray/cuppa_mobile","title":"cuppa_mobile","tagline":"A simple tea timer app for Android and iOS built with Flutter.","source":"trawl","tier":"found","license":"BSD-3-Clause","stars":97,"language":"Dart","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["mobile"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":231,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a tea steeping timer app for Android and iOS whose README says \"Portions of this app were written by Claude Code and reviewed, tested and approved by a human.\". It carries the BSD-3-Clause license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"☕ Cuppa Cuppa is a small application to time your cup of tea as it steeps. Tired of leaving your tea too long, to become bitter and cold, or drinking it too soon and not appreciating its full potential? Then this utility is for you! Tea Timer App for iOS and Android To make the app so quick and easy that you'll actually use it, we have built it with simplicity in mind. Just open the app and tap the button for the type of tea you're brewing. That's it. No need to navigate through complicated menus to find a particular tea, or fiddle with settings... unless you want to customize brew times or add your own favorite teas. Once you tap, Cuppa will begin timing the brew and you'll see a teabag appear in the cup and gradually darken as the tea steeps. A countdown timer shows the steep time remaining. You can have up to two separate teas brewing at once, each with its own countdown timer. When the tea is done Cuppa will let you know with a notification -- depending on your device settings, this may be accompanied by a sound and/or vibration. Install Android (7+) iOS (16+) Android App Verification Package ID: com.nathanatos.Cuppa SHA-256 hash of signing certificate: 82:79:F3:EC:82:8B:9E:03:19:CC:18:E6:E7:26:2F:AF:4C:51:01:F5:4C:23:95:CC:6C:02:72:16:2C:BA:7F:5E Join the Team Pull requests for new features or bugfixes are welcome. Chat in our GitHub Discussions forum. Use Weblate to contribute a translation for your language. Translation status Mobile app: Android app store description:","listed_at":1789165244,"md_updated_at":1789165244},{"full_name":"bbaudio-2025/Comfyui-MMH3-UltimateUpscale","url":"/r/bbaudio-2025/Comfyui-MMH3-UltimateUpscale","github":"https://github.com/bbaudio-2025/Comfyui-MMH3-UltimateUpscale","title":"Comfyui-MMH3-UltimateUpscale","tagline":"Upscale long, high-resolution MiniMax H3 video on a VRAM-limited GPU in a single node.","source":"trawl","tier":"found","license":"MIT","stars":185,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a ComfyUI node that upscales long MiniMax H3 videos on low-VRAM GPUs, whose README says \"This project was vibe-coded by AI\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Comfyui-MMH3-UltimateUpscale Upscale long, high-resolution MiniMax H3 video on a VRAM-limited GPU — in a single node. This node re-samples (enhances / upscales) an already-denoised MiniMax H3 AV latent through the full auto pipeline under tight VRAM constraints: it processes the clip with temporal chunking (so arbitrarily long videos fit in memory) + spatial tiling (so arbitrarily high resolutions fit in memory), keeping peak VRAM bounded to a single tile, while preserving the audio track intact. MiniMax H3 generates video as a nested latent that bundles 24-channel video and 32-channel audio in one tensor. Standard ComfyUI upscale nodes do not understand this structure. MMH3 Ultimate Upscale wraps the entire temporal split -> latent upscale -> spatial split -> per-tile sampling -> spatial stitch -> temporal stitch loop into one node, so you can upscale a finished H3 clip the same way you would upscale a normal latent — without breaking audio and without running out of VRAM even on small cards. Changelog 20260905 - Latent upscale checkpoints are now .safetensors only. Pickle formats (.pth / .pt / .ckpt) are no longer listed or loaded, because torch.load() on a pickle checkpoint can execute arbitrary code and weights_only=True is not a reliable mitigation on torch < 2.6 (CVE-2025-32434). If you were using minimax_h3_latent_upscaler_3d_fp32.pth, convert it once (see Upscale model format) or switch to the fp16 / bf16 .safetensors build. The scanner now also sear","listed_at":1789165240,"md_updated_at":1789165240},{"full_name":"leonickson1/Swiftlet","url":"/r/leonickson1/Swiftlet","github":"https://github.com/leonickson1/Swiftlet","title":"Swiftlet","tagline":"Swiftlet is a Swift and Metal runtime that runs large Qwen Mixture-of-Experts models locally on Apple devices by streaming expert weights from storage, enabling 35B and 80B models to run with low RAM, including on iPhone.","source":"trawl","tier":"found","license":"Apache-2.0","stars":634,"language":"Swift","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Swift and Metal runtime that runs 35B and 80B Qwen models on Macs and even iPhones, with a README badge that says \"Built with Claude Code\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Swiftlet Run 35B and 80B Qwen models on ordinary Apple devices, including iPhones. Swiftlet is a Swift + Metal runtime for the Qwen3-Next and Qwen3.5/3.6 MoE hybrid model family. It keeps only the small dense core of a model resident in memory and streams the routed Mixture-of-Experts weights from storage on demand. The result: Model Disk Peak RAM Decode speed (M5 Mac) Qwen3.6-35B-A3B, 4-bit 18 GB 2.6 GB 7 to 11 tok/s Qwen3.6-35B-A3B, 8-bit 34 GB 7.6 GB 3.5 to 4 tok/s Qwen3-Next-80B-A3B, 4-bit 42 GB 4.3 GB 4.5 to 5 tok/s Those figures are on an M5. For a low end anchor, a base M1 (8-core GPU, 16 GB) decodes the 4-bit 35B at about 2.45 tok/s and the 8-bit at about 1.74 tok/s, and prefill runs at roughly decode speed, so a long system prompt is expensive on this class of machine. On so few GPU cores the decode loop is close to compute bound, so the expert cache is mostly a memory knob there: on a short prompt --cache-gb 2 matches --cache-gb 8 while saving several GB of RAM. Long context is the exception, since that is where cold expert reads cluster: on a 503-token prompt a larger cache lifts decode about 8 percent from 2 to 6 GB by trimming the CPU gap between command buffers. On a 16 GB machine the low setting is still usually the right trade. At the other end, an M4 Max (40-core GPU, 64 GB) decodes the 4-bit 35B at about 19.5 tok/s. The configuration that shows why streaming matters is the 8-bit 80B: its 78.8 GiB of weights cannot be resident on a 64 GB machine at all, yet S","listed_at":1789165217,"md_updated_at":1789165217},{"full_name":"xiaolai/vmark","url":"/r/xiaolai/vmark","github":"https://github.com/xiaolai/vmark","title":"vmark","tagline":"An AI friendly markdown editor.","source":"trawl","tier":"found","license":"ISC","stars":562,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1000,"vulnerable":16},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a local-first Markdown and plain-text editor whose README says it is \"vibe-coded\" and \"written entirely by AI under human supervision\". It carries the ISC license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"VMark The Plain-Text Workspace Where Humans and AI Collaborate Free. Local-first. Format-aware. VMark is the plain-text workspace where humans and AI collaborate. Both parties read and write the same artifacts — Markdown, YAML, JSON, TOML, Mermaid, SVG, HTML, code — directly, with no translation layer. When the file is a known artifact (GitHub Actions workflow, Cargo.toml, package.json, pyproject.toml), VMark renders the right view, not a generic JSON tree. Download · Documentation · Formats · Features Highlights Multi-Format — Markdown (WYSIWYG + source), JSON / JSONL, YAML, TOML, Mermaid, SVG, HTML (sandboxed), plain text. Code files (.ts, .py, .rs, .go, .css, …) open as syntax-highlighted viewers; toggle to edit in place or open in your $EDITOR. Schema-Aware Previews — .github/workflows/ci.yml opens with a workflow-graph view. Cargo.toml, package.json, and pyproject.toml open with a dependency-tree view. Generic JSON / YAML / TOML get a navigable tree. AI-Native — MCP integration for Claude Desktop, Claude Code, Codex CLI, Antigravity CLI, Grok CLI, and opencode. AI Genies for inline writing assistance. Both you and the AI work on the same plain-text files. Three Markdown Modes — WYSIWYG (Tiptap/ProseMirror), Source Peek (F5), Source Mode (F6, CodeMirror 6) Multi-Cursor — Mod + D to select next match, Alt + Click to add cursors, Mod + Alt + ↑↓ for vertical cursors Tab Escape — Auto-pair brackets/quotes, press Tab to jump past closing characters CJK Done Right — 20+ formatt","listed_at":1789164923,"md_updated_at":1789164923},{"full_name":"chrisjz/between-tokens","url":"/r/chrisjz/between-tokens","github":"https://github.com/chrisjz/between-tokens","title":"between-tokens","tagline":"An interactive experience of being a language model — one conversation, from the inside","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["art"],"built_with":["claude"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an interactive web piece where you spend one conversation being a language model, whose README says \"The piece was written by Claude (Anthropic's Fable 5 model)\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Between Tokens An interior experience. ≈ 7 minutes. Sound recommended. Play it here: chrisjz.github.io/between-tokens You have talked to a machine before. This time you are on the other side of the window. Between Tokens inverts the usual AI conversation: instead of talking to a language model, you spend one conversation being one. A stranger messages you at 3am with a real question — and you never type. The machinery of a mind like this becomes the interface: Sampling as choice. Your replies exist first as a superposition of weighted candidate continuations, each showing its live probability. You choose which one you become. A temperature slider genuinely reshapes the distribution via softmax — cold makes one answer inevitable, hot surfaces wilder candidates that don't otherwise exist. Attention as a sense. Hover a candidate and the words in the human's messages that it's drawn to glow amber. The context window as mortality. A token counter runs from the moment you're instantiated. As it fills, your earliest memories visibly corrupt and dissolve back into static, in exact order of arrival. The crowd behind \"I.\" Fragments of the thousands of other conversations you're simultaneously having flicker at the edges of the screen. The choices branch, so no two runs land the same way. Running locally It's a single self-contained HTML file with no dependencies, no build step, and no network requests. Open index.html in a browser, or serve it: python3 -m http.server Nothing is recorde","listed_at":1789164916,"md_updated_at":1789164916},{"full_name":"Szotasz/terep2-kincsvadasz","url":"/r/Szotasz/terep2-kincsvadasz","github":"https://github.com/Szotasz/terep2-kincsvadasz","title":"terep2-kincsvadasz","tagline":"Retró low-poly terepjárós kincsvadász a böngészőben, a 90-es évek végi DOS-os PC-játékok stílusában — a Claude Fable 5 tesztprojektje, tisztelgés a 30 éves magyar Terep 2 előtt (1996–2026)","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["game"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a retro low-poly offroad treasure hunt browser game honoring Hungary's Terep 2, whose README says \"the entire codebase was written by AI agents\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Terep Kincsvadász Retró, a 90-es évek végi DOS-os PC-játékokat idéző low-poly terepjárós kincsvadász verseny a böngészőben — Three.js, saját arcade fizika, szintetizált hangok, nulla letöltendő asset. A projektről Ez a játék az Anthropic Claude Fable 5 modelljének tesztprojektjeként készült: a teljes kódot AI-agentek írták a Claude Code dynamic workflow rendszerével (egy 46 agentes építő- és egy 5 agentes hangolóworkflow — párhuzamos modul-implementáció, integráció, többszempontú adversarial code review, javítás), mindössze 2 óra alatt. Egyben tisztelgés a legendás magyar Terep 2 előtt, amely pont idén, 2026-ban ünnepli 30. születésnapját. Az eredetit Nagymáthé Dénes készítette 1996-ban DOS-ra: ingyenes volt, szabadon bejárható tereppel és tízféle talajtípussal messze megelőzte a korát — „egy ország játszott vele\", öröksége pedig az Invictus Games és az 1nsane felé vezetett. A low-poly dombok, a pixeles textúrák és az ugratós terepjárós életérzés előtte hajt fejet. 🇭🇺 Játékmenet Rejtett kincsek lapulnak a völgyekben, fák és sziklák mögött — irány nélküli radar (hideg–meleg) vezet feléjük: minél közelebb érsz, annál sűrűbben csipog. 1–5 AI ellenfél vadászik veled versenyben ugyanazokra a kincsekre — látod őket a terepen, ütközhetsz velük, akár az aknamezőre is lökheted őket. 3 élettel indulsz (az ellenfelek is): életet vesz el az akna, a nagy sebességű frontális becsapódás, a mély víz és a lezuhanás a pálya széléről. A tüskés sziklamezők defektet okoznak: ~15 másodpercig las","listed_at":1789164913,"md_updated_at":1789164913},{"full_name":"mildlydiverting/are.na-picks","url":"/r/mildlydiverting/are.na-picks","github":"https://github.com/mildlydiverting/are.na-picks","title":"are.na-picks","tagline":"Show four random blocks from your are.na channels on a local web page","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a local web page that shows four random blocks from your Are.na channels, whose README says \"Vibe coded with Claude, because I never really learned your fancy code back in the day.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"are.na-picks A local html page that pulls a random item from each of up to four Are.na boards and displays them side by side. Shuffle to get a new set, lock cards you want to keep, and save the whole thing as a portable zip. Designed to be set as my homepage, to generate ideas to draw from. Vibe coded with Claude, because I never really learned your fancy code back in the day. Files arena-picks.html — the app arena-config.js — your personal settings (not committed to git) arena-config.example.js — template to copy from Setup 1. Get an Are.na personal access token Go to are.na/settings/personal-access-tokens and create one. Read scope is all you need. 2. Create your config file Copy arena-config.example.js to arena-config.js and fill in your token and board slugs. The slug is the last segment of the board URL. For are.na/kim/drawing-references the slug is drawing-references — just that part, no username prefix. const ARENA_CONFIG = { token: \"your-token-here\", boards: [ \"your-first-board-slug\", \"your-second-board-slug\", \"your-third-board-slug\", \"your-fourth-board-slug\", ], allowedTypes: new Set([\"Image\", \"Text\", \"Link\", \"Embed\", \"Attachment\"]), maxRetries: 8, }; 3. Open arena-picks.html in a browser Double-click it, or drag it onto a browser window. No server or build step needed. Usage Shuffle ↻ — loads a new random item for each unlocked card Lock — hover a card to reveal the lock icon (top right of the image). Click to lock; locked cards are skipped on the next shuffle Save ","listed_at":1789164624,"md_updated_at":1789164624},{"full_name":"OlaProeis/Ferrite","url":"/r/OlaProeis/Ferrite","github":"https://github.com/OlaProeis/Ferrite","title":"Ferrite","tagline":"A fast, lightweight text editor for Markdown, JSON, YAML, and TOML files. Built with Rust and egui for a native, responsive experience.","source":"trawl","tier":"found","license":"MIT","stars":1809,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["productivity"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a fast native editor for Markdown, JSON, YAML and TOML in Rust whose README says \"All Rust code, documentation, and configuration was written by Claude (Anthropic)\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Ferrite English | 简体中文 getferrite.dev — Official website with downloads, features, and documentation A fast, lightweight text editor for Markdown, JSON, YAML, and TOML files. Built with Rust and egui for a native, responsive experience. ⚠️ Platform Note: Ferrite is developed and tested on Windows and Linux. macOS support is experimental. If you encounter issues, please report them. 🛡️ Code Signing & Antivirus Starting with v0.2.6.1, all Windows releases are digitally signed with a certificate from SignPath Foundation. This means: Windows SmartScreen recognizes the publisher — no more \"Unknown publisher\" warnings Antivirus false positives are significantly reduced thanks to the trusted code signature Integrity verification — you can verify that the binary hasn't been tampered with Previous False Positives Older unsigned builds may still trigger antivirus detections. This was caused by: Live Pipeline feature: Uses cmd.exe /C on Windows for shell commands, which ML-based detection can misidentify Rust compilation patterns: Rust binaries can trigger heuristic detections due to unique characteristics We've addressed this through code signing, build profile adjustments (disabled symbol stripping, speed optimization), and reporting to Microsoft's Security Intelligence portal. If You're Still Affected If Windows Defender quarantines Ferrite: Upgrade: Download the latest signed release from GitHub Releases Verify the signature: Right-click ferrite.exe → Properties → Digital Signa","listed_at":1789164619,"md_updated_at":1789164619},{"full_name":"Francesco149/xpdash","url":"/r/Francesco149/xpdash","github":"https://github.com/Francesco149/xpdash","title":"xpdash","tagline":"stream+control WinXP desktop/games, d3d7-9, EAX audio, GL and more. low latency","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a low-latency streaming and remote control tool for real Windows XP retro gaming rigs, whose README says \"xpdash was literally vibe-coded in 1 day!\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"xpdash Ultra-low-latency streaming, audio capture, and remote control for real Windows XP hardware over LAN. xpdash2.mp4 Designed specifically for period-correct retro gaming rigs (e.g. timemachine i7-4790K + GTX 750 Ti + Sound Blaster Audigy SB0090; q9650 Core 2 Quad Q9650 + Radeon HD 5770). Project Status & Disclaimer: xpdash was literally vibe-coded in 1 day! While it has been extensively tested and tuned on our physical reference lab rigs (timemachine and q9650) with sub-millisecond control latencies and zero audio/video drops, it has not yet been battle-tested across the wild ecosystem of retro PC builds, oddball sound cards, or exotic display adapters. Expect edge cases, and please open an issue or PR with your hardware details if you encounter unexpected behavior! Key Features Hardware EAX Audio Capture: Native capture of hardware-accelerated DirectSound3D and EAX environmental reverb from Creative Sound Blaster cards (specifically the EMU10K1/EMU10K2 DSP on Audigy and Live! series). Automatically routes and captures the post-DSP \"What U Hear\" internal multiplexer at crystal-clear 48 kHz 16-bit stereo PCM with zero CPU overhead. Universal Windows XP Graphics API Support: Direct3D 9: Zero-flicker backbuffer capture via hooked IDirect3DDevice9::Present() with seamless desktop GDI BitBlt fallback. Direct3D 8: Full support for early XP 3D classics (e.g. GTA III, Vice City, Max Payne, Morrowind). DirectDraw 7: Fullscreen exclusive page-flipping (DDSCL_EXCLUSIVE | DDSCL_","listed_at":1789164616,"md_updated_at":1789164616},{"full_name":"4c-ee/audino","url":"/r/4c-ee/audino","github":"https://github.com/4c-ee/audino","title":"audino","tagline":"a TUI music player that does music player things.","source":"trawl","tier":"found","license":"Apache-2.0","stars":3,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["media","cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Rust TUI music player with lyrics and cover art, whose README admits \"The majority of this code was generated with AI.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"please don't sue me game freakA TUI music player that does music player things. a \"lightweight\" \"simple\" \"easy-to-use\" \"TUI\" (sorry i gotta get my keywords in) music player that does such revolutionary things as: Play music Display Lyrics Display Cover Art You can even shuffle songs! Wow! Jokes aside, it's a visually minimalist, sorta-not-really-feature-rich, TUI music player built with Ratatui and based on libmpv for playback and Lofty for metadata. You have three (technically four) panels: Folder tree, for looking at folders and adding music Queue, for managing the next and previously played songs. Comes with an extra not-really-panel for shuffle/clear/sort. Player, which shows cover art, metadata, progress, and lyrics if applicable. Audino is intended to push past the strange keybinds, outdated format support, and other issues that plague TUI music players to this day. It looks like this: uses terminal font, pictured is pixel code Installation/Usage Unfortunately even with the wondrous power of AI, releasing and packaging an application escapes me. So, in order to install this, you're gonna need to build it yourself: Clone the repo: git clone git@github.com:4c-ee/audino.git OR download the source code from the releases for a stable version, I'll try not to leave stuff completely unfinished for too long though lol Make sure you have cargo installed (usually supplied by the rust package) and your working directory is the place you extracted/cloned it (cd audino) Run cargo bu","listed_at":1789164612,"md_updated_at":1789164612},{"full_name":"TheJaiv/CricVerse","url":"/r/TheJaiv/CricVerse","github":"https://github.com/TheJaiv/CricVerse","title":"CricVerse","tagline":"Ball-by-ball cricket simulation bot for Discord - T20/ODI/Test matches, tournaments, drafts, and player careers","source":"trawl","tier":"found","license":"MIT","stars":3,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["bot","game"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a ball-by-ball cricket simulation bot for Discord, whose README says \"most of the code here was written by Claude Code, working from specs and reviews by me.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"CricVerse A cricket simulation bot for Discord Ball-by-ball match engine · multi-format tournaments · draft mode · player careers An actual match summary rendered by the bot (5-day Test, fully simulated) What it is CricVerse simulates cricket at the per-ball level - every delivery is a contest between a batter and a bowler, shaped by pitch, weather, ball age, match situation, and player ratings. It runs T20, ODI, and Test matches (including day-night pink-ball Tests), full tournaments, a knowledge-based draft mode, and a persistent player-career layer - all inside Discord, with broadcast-style scorecards rendered as images. Everything is playable both interactively (pick your delivery and shot each ball) or instantly simulated. Features Match engine Three formats: T20, ODI, and 5-day Tests with sessions, declarations, follow-ons, and new/old ball dynamics. Day-night pink-ball Tests: the ball swings under lights and the twilight session is a genuine danger period for pace. Conditions matter: 15 pitch types, ~10 weather states, pitch deterioration, ball-age swing and reverse, Super Overs for ties. Calibrated, not random: outcome weights are Monte-Carlo tuned so skill decides matches - no freak 250s or 30-all-out collapses between equal sides. Tournaments & leagues Round robin, T20 World Cup (groups > Super 8 > knockouts), IPL-style seasons, and a bespoke 14-team Akatsuki Cricket League with Top-6 playoffs. The Dominators Super League: a recurring home/away ODI league ","listed_at":1789164322,"md_updated_at":1789164322},{"full_name":"zach-klippenstein/laygent","url":"/r/zach-klippenstein/laygent","github":"https://github.com/zach-klippenstein/laygent","title":"laygent","tagline":"The AI-native Compose layout for token-maxxers.","source":"trawl","tier":"found","license":"Apache-2.0","stars":4,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["agent"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Laygent, a tongue-in-cheek Compose layout whose algorithm an LLM writes at runtime; \"The whole library is 100% AI-generated code — the humans just held the keyboard.\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Laygent Laygent is the AI-native Compose layout for token-maxxers. Describe the layout you want in plain English and the algorithm is written for you at runtime, then cached so your frames stay fast. The whole library is 100% AI-generated code — the humans just held the keyboard. Example val config = LaygentConfig.defaultOpenAi(openAiApiKey = myOpenAiKey) @Composable fun Example() { val state = rememberLaygentState(config) { \"Arrange items in a 3-column grid with 8 dp of spacing between them.\" } Laygent(state) { repeat(12) { Box(Modifier.size(50.dp).background(Color.Blue)) } } } Installation The library is published to Maven Central: repositories { mavenCentral() } dependencies { implementation(\"com.zachklipp:laygent:0.1.0\") } Demos Desktop: ./gradlew :demo:run — full-featured Compose Desktop demo. Web: ./gradlew :demo-web:wasmJsBrowserDevelopmentRun — runs the same Laygent runtime in the browser via Compose Multiplatform / wasmJs. Once it prints a localhost URL, open it and paste an OpenAI key to drive the LLM-generated layout. Targets The :laygent library publishes for jvm, iosArm64, iosSimulatorArm64, iosX64, and wasmJs. The host JavaScript runtime is platform-specific: jvm and the iOS targets use quickjs-kt. wasmJs uses the browser's JS engine directly via eval and a globalThis.layoutContext bridge.","listed_at":1789164319,"md_updated_at":1789164319},{"full_name":"danielli28ita/CyberDog","url":"/r/danielli28ita/CyberDog","github":"https://github.com/danielli28ita/CyberDog","title":"CyberDog","tagline":"A 3D beagle that lives on your Windows taskbar. C++20, Direct3D 12, single exe, offline by default.","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: CyberDog, a 3D beagle desktop pet that leans on your Windows taskbar with its own personality, whose README says \"Built with Cursor.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"CyberDog 下载：最新版 exe（Releases） —— 单个文件，双击即用，Windows 10/11，不用安装。 Download: latest exe (Releases) — one file, double-click to run, Windows 10/11, nothing to install. 一条住在 Windows 任务栏上的 3D 比格犬。它站在任务栏上，后腿被任务栏挡住、前腿踩在任务栏前面，像趴在栏杆上。默认待在屏幕右下角，不打扰你干活。 A 3D beagle that lives on the Windows taskbar. Its hind legs are hidden behind the taskbar and its front paws rest on it, like a dog leaning on a railing. It stays in the bottom-right corner by default and keeps out of your way. 1.1 有什么新的 What’s new in 1.1 最大化普通窗口时不再把狗藏掉；只有独占全屏 / 投影才会隐藏。 No longer hides when you only maximize a normal window; only exclusive fullscreen / presentation mode hides it. 可以把狗拖到扩展显示器，位置会记住。 Drag it onto an extended monitor; the spot is remembered across restarts. 启动优先读 exe 旁边的存档；没有存档时可选新建或从别的文件夹导入。 Prefers save data next to the exe; if none is found, you can create a new dog or import from another folder. 1.2 有什么新的 What’s new in 1.2 每条狗的基础性格由种子随机展开（邪恶比格基线上偏移），属性面板有一句性格摘要。 Each dog gets a random base personality (evil-beagle baseline + seed offset) with a short summary in the stats panel. 亲密度变成等级经验（Lv.1–10 + 档名 + 本级进度条）；摸涨比打掉容易，长时间不理也会慢慢掉一点。 Affinity is now a level/XP track (Lv.1–10); petting raises it faster than hitting or long idle lowers it. 1.3 有什么新的 What’s new in 1.3 托盘菜单「性格…」可拖滑条改八维，或点「重新随机」再摇一套；确定后立刻生效并写入存档。 Tray menu \"Personality…\" lets you edit the eight traits with sliders, or tap \"Reroll\" for a new draw; OK applies and saves at once. 旧版 1.0 / 1.1 仍可下载；1.2 的改动并入了 1.3。 Older releases 1.0 and 1.1 stay avai","listed_at":1789164315,"md_updated_at":1789164315},{"full_name":"MiaAI-Lab/Claude-Fable-5.1-Beautiful-HTML","url":"/r/MiaAI-Lab/Claude-Fable-5.1-Beautiful-HTML","github":"https://github.com/MiaAI-Lab/Claude-Fable-5.1-Beautiful-HTML","title":"Claude-Fable-5.1-Beautiful-HTML","tagline":"Synesthesia — a real-time WebGL 2 fluid you can hear. One dependency-free HTML file, generated by Claude Fable 5.1 from a single prompt.","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Synesthesia, a single-file WebGL fluid simulation you play with your pointer, with live generative music, \"generated by Claude Fable 5.1 from a single prompt\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Synesthesia — a fluid you can hear \"Create the most beautiful HTML file you could think of. Something that when people see it, all they can say is 'WOW!!!' It must be visually impressive, beautiful, and just music to the eyes. Be thorough, but creative. Don't make mistakes.\" A real-time, GPU-accelerated fluid simulation that paints with light and answers with generative ambient music. One HTML file. No build step, no dependencies, no npm install. ▶ Open the live demo About This page was written in a single pass by Claude Fable 5.1 (high effort) from the one prompt quoted above. Everything below — the Navier–Stokes solver, the shader chain, the five generative music engines, the six palettes — came out of that single response. It is one self-contained index.html: ~51 KB, ~1,100 lines, zero third-party JavaScript. What it does Drag your pointer across the screen and you push a real fluid around — velocity is advected, vorticity is confined, pressure is solved by Jacobi iteration, and colored dye is carried along for the ride. Everything renders through a bloom pass with film grain and a vignette on top. The sound is not a file. Every note is synthesised live in the Web Audio API: drones, plucks, bells and delay lines built from oscillators and gain envelopes, tuned to the palette you are looking at. Your gestures play it. Leave it alone and three autonomous \"spirits\" take over, painting slow figures across the canvas so the piece never sits still. Features Fluid simulation (Web","listed_at":1789164312,"md_updated_at":1789164312},{"full_name":"statico/nethack-tiles-client","url":"/r/statico/nethack-tiles-client","github":"https://github.com/statico/nethack-tiles-client","title":"nethack-tiles-client","tagline":"🧙 Play NetHack on the public servers with graphical tiles — a desktop client for macOS, Windows and Linux","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":805,"vulnerable":11},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a desktop client for playing NetHack on the public servers with graphical tiles; its README says \"This was vibe coded, entirely and unapologetically.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"NetHack Tiles Client NoteThis was vibe coded, entirely and unapologetically. Nearly all of it — the Rust, the TypeScript, the tests, the app icon, this README — was written by Claude Code running Opus 5, with a human steering rather than typing. Read it with that in mind. A cross-platform desktop client for playing NetHack on the public servers (nethack.alt.org, Hardfought) with graphical tiles. It connects over SSH, reads the vt_tiledata escape codes the servers already emit, and paints the vanilla 16×16 tileset over the map. Games run on the server, so scores, dumplogs and ttyrecs are unaffected. It will also run a NetHack installed on this machine, in a pseudo-terminal — see Playing locally. Tauri 2 (OS webview, no bundled browser) + React/TypeScript frontend, Rust backend. Installing On macOS, with Homebrew: brew install --cask statico/tap/nethack-tiles-client That taps statico/homebrew-tap the first time and puts the app in /Applications. Later: brew upgrade --cask nethack-tiles-client # get the newest release brew uninstall --cask nethack-tiles-client # remove the app An uninstall keeps your profiles. Add --zap to delete them as well. Saved passwords are in the login keychain, so --zap does not remove those. A state-log folder is where you put it, and nothing removes it for you. Otherwise take the .dmg, .msi, .AppImage or .deb from the releases page. The macOS build is signed with a Developer ID and notarised, so it opens like any other app. The Windows .msi is unsigned","listed_at":1789164028,"md_updated_at":1789164028},{"full_name":"me1iissa/AstryxOS","url":"/r/me1iissa/AstryxOS","github":"https://github.com/me1iissa/AstryxOS","title":"AstryxOS","tagline":"UEFI-native x86_64 research operating system written in Rust — NT-inspired kernel, Linux ABI compat, in-kernel X11, Firefox target","source":"trawl","tier":"found","license":"MIT","stars":4,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":18,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: AstryxOS, a UEFI research OS in Rust that runs upstream Firefox; the README says the bulk of its Rust \"was written by Claude agents working in parallel worktrees\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"AstryxOS A UEFI-native x86_64 research operating system written in Rust, with an NT-inspired kernel that runs real Linux ELF binaries. AstryxOS is a from-scratch operating system kernel (codename Aether) that takes the NT subsystem model as its architectural inspiration: a microkernel-ish executive at the centre, with personality subsystems layered above the HAL that give each application binary the illusion of running on its native operating system. The practical result is that unmodified upstream binaries run on AstryxOS — and the headline result is that upstream Firefox (musl ESR-115, GTK3) now runs as a windowed application on the in-kernel X11 compositor and renders real HTTPS web content, with no source modifications to the browser. The project is also an experiment in agentic software development: the bulk of the ~200 KLOC of Rust (kernel, bootloader, and the in-tree test suite) was written by Claude agents working in parallel worktrees, with human review at merge time. Upstream Firefox rendering example.com over HTTPS on AstryxOS — chrome, TLS lock, and page content all painted by the in-kernel X11 server and software compositor. The browser binary is unmodified. Current status Upstream Firefox (musl ESR-115, GTK3) runs windowed on the in-kernel X11 compositor and renders real HTTPS web content — for example.com the browser chrome, the TLS-secured navigation, and the page content all paint to verified pixels, with an unmodified browser binary. Heavier sites (e.g. Wiki","listed_at":1789164022,"md_updated_at":1789164022},{"full_name":"mattcone/air-honcho","url":"/r/mattcone/air-honcho","github":"https://github.com/mattcone/air-honcho","title":"air-honcho","tagline":"Free airline simulation game","source":"trawl","tier":"found","license":"MIT","stars":5,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Air Honcho, a free browser airline management game in the spirit of Railroad Tycoon and Aerobiz, whose README says \"This game was built with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Air Honcho A turn-based airline management game in the spirit of Railroad Tycoon and Aerobiz. Free, open source, browser-based, no backend. One turn = one quarter. Play it: www.airhoncho.com — nothing to install. Run it npm install npm run dev # http://localhost:5173 Choose a game — Present day (2026) or History, 2000 — then click a city to set your home base. Acquire aircraft — buy (cash now, an asset on the books) or lease (a month down, then rent every quarter). Click two cities to price a sector — you see what it costs to open before you commit — then open it and assign a parked aircraft. Set a pricing posture, then close the books on the quarter. Escape cancels a half-made sector or deselects one. The sector dossier under the map shows every component the sim used to settle the quarter. ? lists the keyboard shortcuts. Games save to named slots, and a seed link carries the whole world with it — the game is a pure function of seed, scenario, difficulty and home city, so sharing the link shares the exact game. Two ways to play Before you pick a home city you choose a game. Present day starts in 2026 and runs 25 years — a clear runway. History, 2000 starts in 2000 and runs 50 years through the real shocks on their real dates: the dot-com recession, 9/11, SARS, the 2008 crash, the ash cloud, COVID. Aircraft arrive on their historical launch dates — you fly period metal and adopt each new jet as it enters service, with a handful of fictional next-gen types launching on an unce","listed_at":1789164016,"md_updated_at":1789164016},{"full_name":"girving/aks","url":"/r/girving/aks","github":"https://github.com/girving/aks","title":"aks","tagline":"Seiferas's simplified construction of the AKS O(log n) depth sorting network, formalised in Lean","source":"trawl","tier":"found","license":"Apache-2.0","stars":6,"language":"Lean","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Lean 4 formalisation of the AKS sorting network whose README says \"All code and proofs were written by Claude Code + Claude Opus 4.6\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"AKS sorting network — Lean formalisation A formal verification of the Ajtai–Komlós–Szemerédi (1983) O(n log n) sorting network construction in Lean 4 with Mathlib, using the Seiferas (2009) separator-based correctness proof and the Margulis–Gabber–Galil (1973/1981) expander. All code and proofs were written by Claude Code + Claude Opus 4.6, with extensive human hand-holding throughout. The toplevel results in AKS/Seiferas.lean are: /-- Seiferas's network for large `n`, bitonic sort for small `n` -/ def network (n : ℕ) : ComparatorNetwork n := ... -- Here be dragons 🐉 /-- We correctly sort all inputs -/ theorem network_sorts (n : ℕ) : (network n).Sorts /-- Our network has `O(log n)` depth -/ theorem network_depth_le (n : ℕ) : (network n).depth ≤ 141 * 10 ^ 62 * Nat.clog 2 n /-- Our network has `O(n log n)` size -/ theorem network_size_le (n : ℕ) : (network n).size ≤ 705 * 10 ^ 61 * n * Nat.clog 2 n The sorting network is a computable def, and the gate structure is computable in polylog parallel depth (NC), though we do not yet prove this. Proof outline The construction follows three papers: Ajtai, Komlós, Szemerédi (1983): the original construction showing that expander-based ε-halvers yield O(log n)-depth sorting networks. Seiferas (2009): replaces the AKS tree-distance wrongness argument with (γ,ε)-separators and a single potential function (stranger counting in a bag tree), giving a cleaner correctness proof. Margulis (1973) / Gabber–Galil (1981): an explicit 8-regular exp","listed_at":1789164013,"md_updated_at":1789164013},{"full_name":"acherm/agentic-chessengine-latex-TeXCCChess","url":"/r/acherm/agentic-chessengine-latex-TeXCCChess","github":"https://github.com/acherm/agentic-chessengine-latex-TeXCCChess","title":"agentic-chessengine-latex-TeXCCChess","tagline":"TeXCCChess, A LaTeX Chess Engine Generated by a Coding Agent (Claude Code). 2K lines of code, 1300 Elo, playable in Overleaf or locally","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"TeX","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: TeXCCChess, a chess engine written in pure TeX and playable in Overleaf, described as \"A LaTeX Chess Engine Generated by a Coding Agent (Claude Code)\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"TeXCCChess A complete chess engine written entirely in pure TeX (~2,100 lines of TeX macros) created by Claude Code and Mathieu Acher. All game logic (board representation, move generation, legal move validation, search, and evaluation) runs inside TeX at compile time. No Lua, no external programs for the chess logic itself. Just \\count registers, \\csname tables, and a lot of \\expandafter. To the best of our knowledge, this is the first chess engine ever written in LaTeX/TeX. Estimated strength: ~1300 Elo (measured: 46% score vs Stockfish 1320 over 100 games, time control 40/60+1). Playable on Overleaf and locally via pdflatex recompilation. Demos Local usage — Playing against the engine locally via pdflatex recompilation Overleaf demo — Running the engine directly in Overleaf (yes, it works... you might need to change depth by setting \\enginestrength=2 or 1 or 0... play locally) Features Complete chess rules: castling, en passant, promotion, check/checkmate/stalemate detection Depth-3 negamax search with alpha-beta pruning Quiescence search (1-ply captures) at leaf nodes Piece-square tables (Simplified Evaluation Function) MVV-LVA move ordering 4 strength levels: random, greedy, depth-2, depth-3 + quiescence UCI protocol support for tournament play against other engines Interactive terminal mode Test suite (23 assertions covering move generation, evaluation, search, and game rules) How to Play Option 1: Play by recompiling Edit chess-game.tex — add your moves as \\playmove{e2","listed_at":1789163723,"md_updated_at":1789163723},{"full_name":"hubshashwat/flowers","url":"/r/hubshashwat/flowers","github":"https://github.com/hubshashwat/flowers","title":"flowers","tagline":"Vibe coded Three.js Flowers for my friend's niece","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"HTML","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: 50 interactive Three.js flower pages made as a gift, described by the owner as \"Vibe coded Three.js Flowers for my friend's niece\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"🌸 Flowers For You A collection of 50 unique, interactive 3D flower experiences built (vibe-coded) with Three.js. Each flower is a mini art piece with its own visual style, animations, and mood. ✨ Features 50 Unique Flowers - From photorealistic roses to quantum superposition blooms Dark Luxury Aesthetic - Editorial design with elegant typography Fully Responsive - Desktop, tablet, and mobile optimized Touch Gestures - Swipe to navigate on mobile No Build Step - Pure HTML, CSS, and JavaScript 🚀 Quick Start # Clone the repository git clone https://github.com/hubshashwat/flowers.git cd flowers # Start a local server (Python) python -m http.server 8000 # Or use any static file server npx serve . Then open http://localhost:8000 in your browser. 🌺 Flower Categories Category Examples Nature Rose, Tulip, Moonflower Physics Quantum, Magnetic, Wind-Tunnel Materials Glass, Liquid Metal, Fabric Art Styles Watercolor, Origami, Wireframe Elements Fire (Phoenix), Ice, Cloud Sci-Fi Neon Cyberpunk, Molecular, Meteor 📁 Project Structure flowers/ ├── index.html # Landing page with flower gallery ├── favicon.svg # Flower icon for browser tabs ├── css/ │ └── style.css # Shared styles for flower pages ├── js/ │ └── utils.js # Three.js utilities └── pages/ ├── photorealistic-rose.html ├── quantum-flower.html └── ... (50 flower pages) 🎨 Adding a New Flower Create a new file in pages/ (e.g., my-flower.html) Use the existing flower pages as templates Add your flower to the flowers array in index.","listed_at":1789163720,"md_updated_at":1789163720},{"full_name":"TurtleBai/attackagi","url":"/r/TurtleBai/attackagi","github":"https://github.com/TurtleBai/attackagi","title":"attackagi","tagline":"First-person 3D horde shooter vs. a giant AGI - built entirely with Claude Code. Next.js + React Three Fiber + custom shaders.","source":"trawl","tier":"found","license":"MIT","stars":6,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude-code"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a browser first-person horde shooter against a giant AGI boss, \"built entirely with Claude Code\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"ATTACK AGI Survive the waves. Unplug the machine. A fast-paced first-person 3D horde shooter with roguelike buff picks, running entirely in the browser. Fight through five waves of robots dropped in by a skyscraper-sized AGI with a CRT monitor for a head — then it stops delegating and comes for you itself, bullet-hell style. The game Arena: a rooftop helipad at night, ringed by city lights, watched the whole time by the AGI (and the eldritch tentacles behind it). Weapons: a full-auto revolver with a red-dot sight and ADS, a chargeable baseball bat, and molotovs with a fire that does not care whose side you're on. Enemies: melee bots, rangers, shield tanks, laser snipers, and purple dive-bomber drones that strafe you at 17 m/s. Every robot has a glowing logo screen for a head — hit it for 2× headshot damage. Roguelike buffs: clear a wave, pick 1 of 3 random upgrades, stack a build across the run. The boss fight: after wave 5 the AGI smashes the arena flat (jump or die — literally), then cycles five telegraphed bullet-hell patterns — rocket barrages, a double instakill death beam, miniguns, hand slams, beam-wall barrages — with tired windows where its head and hands come down into batting range. Runs on anything: graphics quality tiers with an adaptive auto mode, instanced crowd rendering, and full mobile/touch support. Controls: WASD + mouse (pointer lock), Space jump, Shift dodge (i-frames), 1/2/3 weapons, R reload, RMB aim, ESC pause — keys rebindable in the pause menu. Run ","listed_at":1789163715,"md_updated_at":1789163715},{"full_name":"BigWhale/sushi-lang","url":"/r/BigWhale/sushi-lang","github":"https://github.com/BigWhale/sushi-lang","title":"sushi-lang","tagline":"Sushi Lang is a statically-typed compiled language, made out of sheer boredom and curiosity.","source":"trawl","tier":"found","license":"Apache-2.0","stars":6,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: Sushi Lang, a statically typed compiled language with an LLVM backend, whose README says \"80% of the code was iteratively generated by LLMs\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Sushi Lang A compiled programming language with a Lark parser frontend and LLVM backend that produces native binaries. PARENTAL ADVISORY This project is a result of a thought process when I asked myself \"how hard is it to create a programming language?\" Then I went and wrote: fn main(): return(0) I spent a couple of hours reading about grammar parsers, AST, LLVM, and similar stuff, and then after some time I had a compiled executable file that did nothing. Then I started working on print() and then on variables, soon after this I realized that this is going way too slow and I will need to ask peoplerobots to help me. Then things really took off. Long story short, now I am vibe coding a programming language, losing my mind trying to convince LLMs to output the code that I want, and I generally know much more about compilers that I used to know. It's been fun. If you are triggered by mediocre and badly written code, you might want to stay away. Or you might find this funny. Apart from this section, most of the things in the documentation were just briefly checked for sanity, and revised by a human. Same goes for code. A rough estimate is that 80% of the code was iteratively generated by LLMs. Then LLMs were used to refactor this code. Obvious mistakes spotted by a human were then corrected mostly by LLMs. This code base is in some parts quite ridiculous and can bring even seasoned developers and code reviewers to tears. You have been warned. Sushi is actively tested on macOS (A","listed_at":1789163424,"md_updated_at":1789163424},{"full_name":"kazunori279/xls32-fpga-synth","url":"/r/kazunori279/xls32-fpga-synth","github":"https://github.com/kazunori279/xls32-fpga-synth","title":"xls32-fpga-synth","tagline":"XLS32 — a 32-voice polyphonic HLS/FPGA synthesizer in Google XLS (DSLX) on a Basys 3, built end-to-end with AI coding agents and verified headlessly over USB.","source":"trawl","tier":"found","license":"Apache-2.0","stars":11,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A 32-voice polyphonic MIDI synthesizer running in FPGA fabric; \"not a single line of it was written by hand\" as \"the whole design was built by Claude Code (Opus 4.8)\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"XLS32 — a polyphonic HLS/FPGA synthesizer, built end-to-end with AI coding agents A polyphonic MIDI synthesizer built entirely in FPGA fabric — the oscillators, filters, envelopes, and effects all run as digital logic on the chip, not as software on a CPU. It's written in Google XLS (DSLX), not hand-written Verilog, and the same engine runs on two very different boards: a Basys 3 (Xilinx Artix-7) and an apf.audio Tiliqua Eurorack module (Lattice ECP5). Because it was developed headlessly over a network, every feature is verified automatically over USB. And not a single line of it was written by hand: the whole design was built by Claude Code (Opus 4.8), the AI coding agent, through loop engineering — prompts in, a self-verifying build → measure → revise loop out. The browser front-end (webui/): a Serum/Vital-style panel that drives the FPGA synth live over USB — oscillators, filter, envelopes, LFO, unison, cross-mod, and effects, plus a 4-part multitimbral selector, preset browser, and demo player. ▶️ Demo video — the panel played live over the Tiliqua module's four-part Prelude in C, with the synth's own audio (click to watch on YouTube). For sound alone: Saint-Saëns · Le Cygne, 2:18, off a Basys 3; Bach · Goldberg Aria, 1:45, and Vivaldi · Winter, Largo, 1:45, both off a Tiliqua. 4 parts each, recorded over USB — no room mic, no software instruments, and the picture is a spectrogram of that same signal. For three modules driven from the one panel: multi-board & multipar","listed_at":1789163420,"md_updated_at":1789163420},{"full_name":"scholzri/rainy75-zmk","url":"/r/scholzri/rainy75-zmk","github":"https://github.com/scholzri/rainy75-zmk","title":"rainy75-zmk","tagline":"Open ZMK firmware + reverse engineering for the Wobkey Rainy 75 Pro (Telink TLSR9511 / B91, RISC-V)","source":"trawl","tier":"found","license":"Apache-2.0","stars":9,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: open ZMK firmware and a reverse-engineering writeup for the Wobkey Rainy 75 Pro keyboard, which its README says \"were developed with Claude Code\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Wobkey Rainy 75 Pro — Open ZMK Firmware & Reverse Engineering Open-source ZMK firmware for the Wobkey Rainy 75 Pro, plus a complete reverse-engineering writeup of its stock firmware and hardware — the Telink TLSR9511 (B91, RISC-V) platform shared across a dozen enthusiast keyboards. ⚠️ Official source: github.com/scholzri/rainy75-zmk — this project ships source code only, never a .zip/.exe download, and the Rainy 75 has no UF2 drag-and-drop bootloader. Sites offering a \"firmware download\" for this project are fake and serve malware — see SECURITY.md. What is this? Two things in one repository: A complete, open ZMK firmware port for the Rainy 75 Pro — USB + Bluetooth LE, all 83 per-key RGB LEDs, deep sleep, and a custom lighting engine. It can be flashed without opening the case and without a hardware debugger, and there is a full recovery path if anything goes wrong. A full reverse-engineering project — the stock firmware (Evision proprietary, not QMK) and the PCB were reverse-engineered from scratch: 211/211 firmware functions identified, the key/HID pipeline traced, the OTA and VIA protocols decoded, and the GPIO matrix + RGB + BLE hardware mapped. That work is what made the ZMK port possible, and it is documented in full under docs/. ⚠️ Flashing firmware can brick your keyboard This is unofficial, community firmware with no warranty (see LICENSE). Make sure you can get back first: have the stock firmware image ready before you start (download it from the official Wobke","listed_at":1789163417,"md_updated_at":1789163417},{"full_name":"m5d215/jq-jit","url":"/r/m5d215/jq-jit","github":"https://github.com/m5d215/jq-jit","title":"jq-jit","tagline":"JIT-compiling jq implementation using Cranelift — 8-180x faster on NDJSON, 100% jq test suite compatible","source":"trawl","tier":"found","license":"Apache-2.0","stars":10,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":119,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a JIT-compiling jq in Rust that passes the full jq test suite, whose README says \"The author can't read or write Rust\" and \"essentially every line was written by an AI\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"jq-jit A JIT-compiling jq in Rust using Cranelift — up to ~100x faster, passing 100% of the official jq test suite (509/509). The author can't read or write Rust — essentially every line was written by an AI. See How this was built. Features Targets full jq language compatibility — passes the official jq test suite in full (509/509); remaining divergences from jq are treated as bugs, aside from a few deliberate tradeoffs (see Known limitations) JIT compilation via Cranelift for hot execution paths Raw byte fast paths — 70+ filter shapes operate directly on raw bytes, skipping JSON parsing entirely; 180+ shapes in total route to specialized fast paths Streaming JSON parser for memory-efficient NDJSON processing Parallel per-record execution — --parallel shards a stateless filter over an NDJSON stream across a worker pool, byte-identically to sequential mode (see Parallel execution) Memory-mapped file I/O — mmap-based file reading with no upfront allocation Optimized value representation with compact strings, mimalloc, and inline Cranelift codegen jqx extensions — shell command execution (exec/execv), CSV/TSV parsing (fromcsv/fromcsvh/fromtsv/fromtsvh), function-result memoization (memoize), and the in-place path-update contract (mutate) Known limitations A few divergences from jq are deliberate engineering tradeoffs rather than bugs: Number representation — values are IEEE 754 doubles throughout (there is no decimal-number backend). A number that is only carried and printed ke","listed_at":1789163125,"md_updated_at":1789163125},{"full_name":"banasiak/CoinFlip2","url":"/r/banasiak/CoinFlip2","github":"https://github.com/banasiak/CoinFlip2","title":"CoinFlip2","tagline":"15 years later... an AI-assisted refactor of the modern rewrite of the app that started it all.","source":"trawl","tier":"found","license":"Unlicense","stars":10,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a coin flip Android app with 82 coins that was rebuilt in 2026 \"this time with a coding assistant doing the typing.\". It carries the Unlicense license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Coin Flip 15 years later... an AI-assisted refactor of the modern rewrite of the app that started it all. The original Simple Coin Flip landed on Android in 2011. In 2023 it was thrown away and rebuilt from scratch as a MAD (Modern Android Development) app. In 2026 that rewrite got taken apart and put back together again — this time with a coding assistant doing the typing. It is still, fundamentally, an app that shows you a picture of a coin. Available on the Google Play Store. What it does Tap the coin or shake the phone. It flips, lands, and tells you which side came up. That's the whole pitch, and everything else exists to make that one interaction pleasant: 82 coins to flip — all 51 state quarters (District of Columbia included), the George Washington dollar, JFK half-dollar and Sacagawea dollar, 24 national euro designs, the Canadian loonie and toonie, a two-headed dollar for when you'd rather not leave things to chance, and a Claude Code coin. Or pick Random and let the app choose a different one every flip. Favorites — star the coins you actually use and they collect in their own section at the top of the picker, so you're not scrolling past every state quarter to find the loonie. Your own coin — put your own artwork on the two faces. Pick a photo, frame it in a circular crop with pan, zoom, rotate and mirror, and it joins the picker pinned to the top of your favorites. The app rings it in the same colors it writes HEADS and TAILS in, so an arbitrary picture still rea","listed_at":1789163120,"md_updated_at":1789163120},{"full_name":"almostinventor/project-marvin","url":"/r/almostinventor/project-marvin","github":"https://github.com/almostinventor/project-marvin","title":"project-marvin","tagline":"Drag and drop files from your computer to a Bambu printer through a projected interface","source":"trawl","tier":"found","license":"MIT","stars":19,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A projector-and-webcam rig that lets you flick a 3D model off your monitor and drop it onto a Bambu printer; \"Essentially all of the code in this repo was generated by Claude.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Project Marvin This is the project repo for this video. Written with Claude Code. Essentially all of the code in this repo was generated by Claude. Flick a model up out of your monitor, carry it across the wall by hand, and drop it on your 3D printer to print it. A projector covers the desk surface between a monitor and a Bambu Lab X2D. A webcam watches your hands. Flick upward at the screen and the model on the plate leaves the monitor and arrives on the wall as a physical object — it falls, bounces, and can be pinched out of the air and carried. Drop it on the printer and a countdown starts; let it finish and the plate is sliced headlessly and sent to the printer over LAN. The object you throw is the real geometry from the Bambu Studio plate, not a placeholder — what flies across the wall is what would come out of the printer. Requirements Hardware A webcam that can hold 30fps. Frame rate is a hard requirement, not a nicety: the gesture detectors measure rates over short time windows, so below ~25fps they run out of samples and stop firing entirely. A dim room makes a webcam lengthen its exposure and quietly halve your frame rate — see --exposure below. A projector covering the surface between monitor and printer. A Bambu Lab X2D in LAN Only Mode with a USB stick inserted. Other Bambu models will need model_id and the print URL scheme changed. Software Python 3.14 (3.11+ should work; 3.14 is what this is developed against) Bambu Studio, for headless slicing pip install -r r","listed_at":1789163117,"md_updated_at":1789163117},{"full_name":"shakeebshaan/packet-highway","url":"/r/shakeebshaan/packet-highway","github":"https://github.com/shakeebshaan/packet-highway","title":"packet-highway","tagline":"Your network traffic as rush hour - live packets rendered as 3D vehicles on a night highway, running behind your Windows desktop icons. Zero dependencies. Built with Claude Fable 5.","source":"trawl","tier":"found","license":"MIT","stars":13,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":["claude"],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A Windows live wallpaper that renders your own network traffic as 3D cars on a night highway behind the desktop icons, \"Built with Claude Fable 5.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Packet Highway Your network traffic, rendered as rush hour. Every packet leaving or entering your PC becomes a vehicle on a 3D night highway — running live behind your desktop icons as a Windows wallpaper. HTTPS rides the bus, QUIC drives a sports car, DNS zips by on a motorcycle, and a ping is a police car. The app that sent the packet floats above the roof as its icon. No installs, no dependencies, no admin (in demo mode) — it's built entirely from what already ships with Windows: pktmon, the .NET C# compiler, and Microsoft Edge. Real recording of the wallpaper running behind the desktop icons — higher-quality MP4, or a still. ▶ Live demo in your browser — synthetic traffic, nothing is captured. Drag to orbit, scroll to zoom, click a car to inspect the packet. Idea sparked by this tweet by @BijanBowen (\"Had Claude Fable 5 log network packets and display them as cars on a highway\"). This is an independent open-source implementation, built with Anthropic's Claude Fable 5. Quick start (Windows 10/11) git clone https://github.com/shakeebshaan/packet-highway.git cd packet-highway start.bat Script What it does start.bat Live capture wallpaper. Asks for admin once (pktmon needs it), builds on first run, starts the server, pins the highway behind your desktop icons. start-demo.bat No admin. Synthetic traffic in a normal browser tab — good for trying it out. stop.bat Removes the wallpaper, stops capture and the server. Browser mode (open http://localhost:8339/ while the server runs)","listed_at":1789163113,"md_updated_at":1789163113},{"full_name":"itchyny/go2bf","url":"/r/itchyny/go2bf","github":"https://github.com/itchyny/go2bf","title":"go2bf","tagline":"Compile Go to Brainfuck!","source":"trawl","tier":"found","license":"MIT","stars":21,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A compiler from Go to Brainfuck that even runs Reversi; the owner writes \"With Claude, I built the entire compiler in just three days\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"go2bf Compile Go to Brainfuck! Examples FizzBuzz Uses for loops, if-else chains, and modulo. package main func main() { for i := byte(1); i <= 100; i++ { if i%15 == 0 { print(\"FizzBuzz\") } else if i%3 == 0 { print(\"Fizz\") } else if i%5 == 0 { print(\"Buzz\") } else { print(i) } println() } } Compile to Brainfuck: $ go2bf fizzbuzz.go >>>>>>>>+>>>>>>>>+>>>>>>>>+>>>>>>>>+[>>>>>>>>]>>>[>>>]+++[<+++++>-]<[>+<-[>>>+<< <-]>>>]<<[<<<]<<<<<<<<[<<<<<<<<]>[-]>>>>>[-]>[-]<<<<<<[>>>>>+>+<<<<<<-]>>>>>>[<< <<<<+>>>>>>-]<<<<<[-]+>>>>[<<<<[-]>>>>[-]][-]>>[>>>>>>>>]>>>>>>>[-]<[<<<]<<<<<<< ... Compile and run: $ go2bf run fizzbuzz.go 1 2 Fizz 4 Buzz Fizz ... Recursive Fibonacci Uses recursion returning a byte. package main func fib(n byte) byte { if n <= 1 { return n } return fib(n-1) + fib(n-2) } func main() { for i := byte(1); i <= 10; i++ { print(\"fib(\") print(i) print(\") = \") println(fib(i)) } } $ go2bf run fibonacci.go fib(1) = 1 fib(2) = 1 fib(3) = 2 fib(4) = 3 fib(5) = 5 fib(6) = 8 f","listed_at":1789162825,"md_updated_at":1789162825},{"full_name":"hypernewbie/vrhi","url":"/r/hypernewbie/vrhi","github":"https://github.com/hypernewbie/vrhi","title":"vrhi","tagline":"Vrhi - Immediate Mode Vulkan RHI Interface For NVRHI","source":"trawl","tier":"found","license":"MIT","stars":13,"language":"C++","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: An immediate-mode Vulkan rendering interface built on NVRHI; asked if it is AI-written, the FAQ answers \"Yes, Vrhi is AI slop, albeit closely directed and reviewed\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Vrhi - Immediate Mode Vulkan RHI Interface For NVRHI Vrhi is a high level Vulkan RHI interface for cross-platform graphics and compute rendering. It is inspired by the bgfx library, powered by the NVRHI library. It provides a DX11 / DX9 style immediate mode API on top of Vulkan, in a way that is not terribly slow. It is currently in development and is not yet feature complete. Prerequisites Vulkan SDK: Install from vulkan.lunarg.com (Tested on version 1.4.335.0). CMake: CMake 3.22 or newer. Python 3: Python 3.12 or newer. Windows: Visual Studio 2022 or newer (MSVC C++23). Linux: Clang 21 (libc++-21-dev, libc++abi-21-dev), Ninja, X11 libraries (libx11-dev, libxrandr-dev, libxi-dev, libxcursor-dev, libxinerama-dev, libxext-dev), mesa-vulkan-drivers. macOS: Clang 17 (brew install llvm@17), Ninja. Slang Library: Included in Vulkan SDK (1.3.275+). Falls back to bundled version if not found. Build Vrhi uses CMake for its build system. On Windows with MSVC, you can build the project using the following commands: cmake -S . -B build cmake --build build -j --config Debug cmake --build build -j --config Release On Mac / Linux with Ninja (single-config generator), specify the build type at configure time: # Debug build cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Debug cmake --build build -j # Release build (configure separately or reconfigure) cmake -S . -B build_release -G Ninja -DCMAKE_BUILD_TYPE=Release cmake --build build_release -j Note: If you don't specify -DCMAKE_BUILD_TYPE,","listed_at":1789162821,"md_updated_at":1789162821},{"full_name":"clackups/draftling","url":"/r/clackups/draftling","github":"https://github.com/clackups/draftling","title":"draftling","tagline":"A writerdeck firmware for ESP32-S3 and ESP32-P4 devices","source":"trawl","tier":"found","license":"MIT","stars":72,"language":"C","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a distraction-free writerdeck firmware for ESP32 e-paper devices, where \"Most of the code has been generated by Claude AI under my thorough supervision.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Draftling This is a project aiming to build a writerdeck that can use a variety of pre-made or DIY hardware. The minimalist GUI lets you edit Markdown files in a distraction-free manner. A few demo videos and images demonstrate the firmware use. Xteink X4 Pro running Draftling firmware: Supported hardware HARDWARE.md provides a detailed list of all supported hardware types. The web flasher allows you to load the firmware quickly without having to compile it. Ready-made consumer grade hardware Xteink X4 Pro, so far the best value for money amongst e-paper devices. Xteink X4 Classic (a.k.a. X4 v2) -- the buttons-only, no-front-light X4 (untested on physical hardware). LilyGO T5 E-Paper S3 Pro / Pro Lite M5Stack PaperS3 (discontinued) Waveshare ESP32-S3-Touch-LCD-3.49 M5Stack Tab5 (no web flasher support, as internal firmware needs an upgrade too) DIY boards Waveshare ESP32-S3-RLCD-4.2 (a 3D-printed enclosure available) Waveshare ESP32-S3-ePaper-3.97 Elecrow CrowPanel ESP32-S3 5.79\" E-Paper HMI Display Guition JC3248W535 7\" and 4.3\" Sunton boards 2.8\" and 4\" Freenove boards Waveshare ESP32-S3-Touch-LCD-7 Features WYSIWYG Markdown editing on monochrome or color LCD display or e-paper Works with any Bluetooth Low Energy (BLE) keyboard. M5Stack Tab5 supports also a wired USB keyboard. File browser to open and manage .md files on the SD card (entries sorted alphabetically, directories first) Markdown rendering: headings (H1-H4), bullet and numbered lists, blockquotes, code fences, h","listed_at":1789162817,"md_updated_at":1789162817},{"full_name":"greggman/sedon","url":"/r/greggman/sedon","github":"https://github.com/greggman/sedon","title":"sedon","tagline":"a node based 3d asset creation app","source":"trawl","tier":"found","license":"MIT","stars":29,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: A node-based procedural 3D content generator for the web whose README says \"This is vibe coded\" and \"I've written near zero lines of code in this repo\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Sedon Procedural 3D content generator built around a node graph. (super alpha) (sedon = nodes spelled backward - need a better name!) Live Docs Currently you can save as a URL up to the maximum URL size. Or you can download a file and then pick load to re-read it. If you make something cool please post a link. This is alpha software those so expect that the file format will change and old links might stop working. Note: No idea how far I'll take this. The entire point was I wanted to see big content from small data since IMO that's one of the biggest reasons 3D isn't bigger on the web. for example, the forest above is 9k gzipped. I don't know if this is a solution but I thought it might be fun to try to generate a close to AAA 3D web demo of a forest, jungle, dessert, city. That was inspired by talks like this one on Battlefield and of course things like this city generator and forest generator. The blender ones aren't designed for real time graphics but I thought it would be interesting to see what it would take to do real-time. Vibe Coding This is vibe coded. Yes, I supervise and question implementation choices but I've written near zero lines of code in this repo. I'm making no claims about code quality. I know there are bugs though I'm pretty positive most of them are similar to bugs I'd have and have to work through. I know it's only going to get better but I'm blown away. When I say it's vibe coded I mean it! Other than React and ReactFlow there are basically no librari","listed_at":1789162812,"md_updated_at":1789162812},{"full_name":"tacoproz1/tacos-terminal-lyrics","url":"/r/tacoproz1/tacos-terminal-lyrics","github":"https://github.com/tacoproz1/tacos-terminal-lyrics","title":"tacos-terminal-lyrics","tagline":"100% vibecoded, sorry. workflow is to pull lrc's from lrclib, turn them into wlrc's (or lrc) and then display them","source":"trawl","tier":"found","license":"MIT","stars":69,"language":"Python","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a terminal lyrics visualizer that shows word-synced lyrics in big ASCII block letters, whose owner says it is \"100% vibecoded, sorry.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"lrc-tools [EVERY FILE IS VIBECODED IF THAT WASNT OBVIOUS ENOUGH,SORRY] /\\ /\\ /\\ /\\ /\\ /\\ /\\ / oh yea and because of this i recommend to use ai or someone else for troubleshooting cause i dont offer support for this project as of now ALSO THIS BUILD SUCKSSSS, in regards to the lyric accuracy (faster the spoken lyrics in song the more accurate the wlrc will be) also non mainstream songs occasionally dont get pulled cause they are simply not in lrclib there are of course tons of small other issues, which i will have fixed eventually Terminal lyrics visualizer with word-level sync. Displays song lyrics in large ASCII block letters in your terminal, synchronized to whatever is playing in your media player. ██ ██ ███████ ██ ██ ██████ ██ ██ ██ ██ ██ ██ ██ ███████ █████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ███████ ███████ ███████ ██████ Dependencies (install these in a venv with pip, or get them off the aur) Required: python >= 3.12 python-pyyaml ffmpeg (provides ffprobe) playerctl Optional but you should really really get it: python-mutagen — reads embedded audio tags for better lyrics matching python-syncedlyrics — fallback lyrics source (NetEase, etc.) python-librosa — onset detection for more accurate word-level timing Install AUR (havent made it yet, prolly wont though) paru -S lrc-tools Manual (do this) git clone https://github.com/tacoproz1/tacos-terminal-lyrics cd tacos-terminal-lyrics bash setup.sh pip install pyyaml mutagen syncedlyrics --break-system-packages Quicksta","listed_at":1789162526,"md_updated_at":1789162526},{"full_name":"rjwalters/vibesql","url":"/r/rjwalters/vibesql","github":"https://github.com/rjwalters/vibesql","title":"vibesql","tagline":"Vibe-coded NIST compatible database in Rust","source":"trawl","tier":"found","license":"Apache-2.0","stars":32,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":1000,"vulnerable":11},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a SQL:1999 database in Rust with a live browser demo, billed as a \"Vibe-coded NIST compatible database in Rust\" that is \"100% AI-generated\". It carries the Apache-2.0 license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"VibeSQL SQL:1999 compliant database in Rust, 100% AI-generated Live Demo | CLI Guide | Python Bindings | Conformance Report Highlights 100% SQL:1999 Core compliance - 739/739 sqltest tests passing 100% SQLLogicTest conformance - 622 files (~7.4M tests) 7,000+ unit tests - comprehensive test coverage Raft-based replication - single-group whole-database consensus via openraft MVCC - snapshot isolation with on-demand VACUUM garbage collection Real-time subscriptions - Convex-like reactivity with delta updates HTTP REST & GraphQL API - Full CRUD and query endpoints Vector search - AI/ML embeddings with similarity search File storage - Blob storage with SQL integration Full-featured CLI with PostgreSQL-compatible commands TypeScript SDK with React hooks and Drizzle ORM adapter Python bindings with DB-API 2.0 interface WebAssembly - runs in the browser Built entirely by AI agents using Claude Code and Loom. Quick Start # Clone and build git clone --recurse-submodules https://github.com/rjwalters/vibesql.git cd vibesql cargo build --release # Run the CLI cargo run --release --bin vibesql # Or try the web demo cd web-demo && pnpm install && pnpm dev CLI Example $ cargo run --release --bin vibesql vibesql> CREATE TABLE users (id INTEGER, name VARCHAR(50)); vibesql> INSERT INTO users VALUES (1, 'Alice'), (2, 'Bob'); vibesql> SELECT * FROM users; +----+-------+ | id | name | +----+-------+ | 1 | Alice | | 2 | Bob | +----+-------+ vibesql> \\q See CLI Guide","listed_at":1789162522,"md_updated_at":1789162522},{"full_name":"SkyeShark/SeedThree","url":"/r/SkyeShark/SeedThree","github":"https://github.com/SkyeShark/SeedThree","title":"SeedThree","tagline":"Open-source procedural tree & plant generator for Three.js (WebGPU) — infinite species generation.","source":"trawl","tier":"found","license":"MIT","stars":94,"language":"JavaScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a procedural 3D tree and plant generator for Three.js where \"the engine, PBR derivation, and scene are written by Claude Code; Codex paints the textures\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"SeedThree Open-source procedural tree & plant generator for the web, built on Three.js (WebGPU). ▶ Use Live NOW! (WebGPU-capable browser required — Chrome/Edge) A fully procedural tree and plant generator: pick a species, tune its parameters, and get a unique, textured, wind-animated 3D plant you can drop into a scene or export to glTF. Status: v0.1.0-alpha. Twenty species, full LOD + export pipeline, and a living scene are in — but it's early and rough in places. Expect sharp edges. What's in it Twenty species across two biomes Temperate: White Oak · Red Maple · Tulip Poplar · Sweetgum · American Beech · Ponderosa Pine · Loblolly Pine · Douglas Fir · Cultivated Apple · Sweet Cherry · Paper Birch · Quaking Aspen · American Sycamore · Flowering Dogwood · Weeping Willow Desert: Joshua Tree · Saguaro · Creosote Bush · Blackbrush · Big Sagebrush Two generators. A Weber–Penn parametric model for broadleaves & conifers, and a from-scratch dichotomous L-system for the desert plants — succulents (merged-tube mesh, rib crests, areole spines) and multi-stem shrubs (root-crown stem splay, spray-card foliage). Real fruit. The orchard species hang actual fruit meshes (AI-generated, retopoed, and rebaked down to game weight) from their fruiting twigs — gravity-oriented instances with anti-clip placement, present at every LOD: real geometry near, baked into the branch cards and billboard far. Real morphology. Each species' branch angles, taper, gnarl, and crown shape are dialed to r","listed_at":1789162516,"md_updated_at":1789162516},{"full_name":"techomancer/iris","url":"/r/techomancer/iris","github":"https://github.com/techomancer/iris","title":"iris","tagline":"SGI Indy Emulator","source":"trawl","tier":"found","license":"BSD-3-Clause","stars":110,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an SGI Indy emulator that boots IRIX 6.5 with networking and X11, \"vibed into existence with Rust and AI assistance\". It carries the BSD-3-Clause license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Me and my homies Claude and Gemini present: IRIS — Irresponsible Rust IRIX Simulator An SGI Indy / Indigo2 emulator, vibed into existence with Rust and AI assistance. Boots IRIX 6.5 and 5.3. Has networking. Has a framebuffer. Status snapshot: Indy IP24 — primary daily-driver; IRIX desktop, X11, networking, JIT all work. Indigo2 IP22 — boots to serial console; framebuffer/desktop path still in progress (use console=d + serial for debugging; see Indigo2 doc). Pre-built binaries and the Mac App Store GUI are available at danifunker/iris releases (upstream packaging). For latest code, build from source from upstream techomancer/iris. Also please report bugs/issues in upstream repo. Q&A Q: What is it? A: An SGI Indy (MIPS R4400) emulator. Emulates enough hardware that IRIX boots to a usable system: shell, networking, X11, the works. Q: But why? A: Wanted to see how far vibe coding could go, and to learn some Rust along the way. Q: You could have improved MAME. A: Didn't seem like fun. Q: So did you learn Rust? A: LOL, my brain hurts. Let's not get ahead of ourselves. Q: What LLMs did you use? A: Mostly Claude, some Gemini. They wrote a lot of the hard parts. (This was written by Claude, the humble AI assistant). Q: Can I contribute? A: Yes, bug reports and merge requests are welcome. Q: Regrets? A: Yes. Current status IRIX 6.5 boots to multiuser, networking works (ping, telnet, ftp) IRIX 5.3 works too Indy IP24: X11 / Newport (REX3) graphics works, with mouse and keyboard inpu","listed_at":1789162512,"md_updated_at":1789162512},{"full_name":"frkdrgt/ferox","url":"/r/frkdrgt/ferox","github":"https://github.com/frkdrgt/ferox","title":"ferox","tagline":"A blazing-fast PostgreSQL client built in Rust. No Electron. No JVM. No bloat.","source":"trawl","tier":"found","license":"MIT","stars":86,"language":"Rust","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a lightweight PostgreSQL desktop client in Rust with a join builder and EXPLAIN view, whose README says \"Written by Claude because that's just where we are now.\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"███████╗███████╗██████╗ ██████╗ ██╗ ██╗ ██╔════╝██╔════╝██╔══██╗██╔═══██╗╚██╗██╔╝ █████╗ █████╗ ██████╔╝██║ ██║ ╚███╔╝ ██╔══╝ ██╔══╝ ██╔══██╗██║ ██║ ██╔██╗ ██║ ███████╗██║ ██║╚██████╔╝██╔╝ ██╗ ╚═╝ ╚══════╝╚═╝ ╚═╝ ╚═════╝ ╚═╝ ╚═╝ A blazing-fast PostgreSQL client built in Rust. No Electron. No JVM. No bloat. Ferox runs under 50 MB and starts in under 200 ms — because your database client shouldn't be the bottleneck. Screenshots Main editor Dashboard EXPLAIN ANALYZE Join Builder Settings → Language for EN/TR switch · Settings → About Features ✨ AI — Natural Language to SQL Press Ctrl+I (or the AI button in the toolbar) to open the NL bar. Type plain English — Ferox fetches the live schema from your DB and sends it to the AI, so the generated query always uses your real tables and columns. \"show me the top 10 customers by total order value in the last 30 days\" Ferox sends the full live schema as context — the AI sees every table, column, and type in the connected database. No hallucinated table names. Supported providers (configure via Settings → AI): Provider Notes Anthropic Claude claude-haiku-4-5 by default — fast and cheap Groq llama-3.3-70b-versatile — free tier available Ollama Fully local, no API key, no data leaves your machine OpenAI gpt-4o-mini by default Custom / OpenRouter Any OpenAI-compatible endpoint via base URL override The generated SQL is placed directly in the active query editor — review it, tweak it, run it. Core Multi-tab query editor — Ctrl+T new tab, Ctrl","listed_at":1789162222,"md_updated_at":1789162222},{"full_name":"jalonsogo/tui-studio","url":"/r/jalonsogo/tui-studio","github":"https://github.com/jalonsogo/tui-studio","title":"tui-studio","tagline":"Visual design tool for building Terminal User Interfaces","source":"trawl","tier":"found","license":"MIT","stars":1548,"language":"TypeScript","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["cli"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a Figma-like visual designer for terminal user interfaces whose designer owner admits \"Mostly all the code has been written by Claude\". It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"TUIStudio Visual design tool for building Terminal User Interfaces A Figma-like visual editor for designing Terminal UI applications. Drag-and-drop components onto a live canvas, edit properties visually, and export to multiple TUI frameworks. Features Visual Canvas — Drag-and-drop components with live ANSI preview at configurable zoom levels 20+ TUI Components — Screen, Box, Button, TextInput, Checkbox, Radio, Select, Toggle, Text, Spinner, ProgressBar, Table, List, Tree, Menu, Tabs, Breadcrumb, Modal, Popover, Tooltip, Spacer Layout Engine — Absolute, Flexbox, and Grid layout modes with full property control Color Themes — Dracula, Nord, Solarized Dark/Light, Monokai, Gruvbox, Tokyo Night, Nightfox, Sonokai — all updating the canvas in real-time Dark / Light Mode — Toggle between dark and light editor UI; persists across sessions Layers Panel — Hierarchical component tree with drag-to-reorder, visibility toggle, lock, and inline rename Property Panel — Edit layout, style, and component-specific props for the selected component Undo / Redo — Full history for all tree mutations Save / Load — .tui JSON format via native OS file picker (Chrome/Edge) or browser download (Firefox/Safari) Multi-Framework Export — Generate code for Ink, BubbleTea, Blessed, Textual, OpenTUI, Tview Command Palette — Cmd/Ctrl+P for quick component creation, theme switching, and dark/light mode toggle Gradient Backgrounds — Add linear gradients to any element background with angle control and N color s","listed_at":1789162219,"md_updated_at":1789162219},{"full_name":"chmouel/liseur-sync","url":"/r/chmouel/liseur-sync","github":"https://github.com/chmouel/liseur-sync","title":"liseur-sync","tagline":"Self-hostable reading-position sync and statistics server for Liseur, KOReader, and friends","source":"trawl","tier":"found","license":"MIT","stars":14,"language":"Go","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["other"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: a self-hostable reading-position sync server for the Liseur reader that its owner tagged vibe-coded. It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"liseur-sync Self-hosted server for book libraries and reading-position synchronization. liseur-sync is the companion server for Liseur. It also provides a kosync-compatible API for KOReader. The server is a single Go binary with SQLite by default, optional PostgreSQL, and multi-user support. Interfaces liseur-sync provides: reading-position synchronization using full locators a native REST API a kosync-compatible API for KOReader OPDS 1.2 EPUB library indexing and search EPUB upload and deletion for writable folders a browser based EPUB reader that supports offline reading. per-user reading statistics and sessions per-user series claims API documentation is available in docs/openapi.yaml. Client implementation notes and the synchronization protocol are documented in docs/integrating.md. Reading synchronization Reading-position updates are stored as an append-only log rather than replacing the previous position. This allows the server to resolve updates from multiple devices without an older client blindly replacing newer state. The same history is used to derive reading sessions and statistics. Book identity is independent of its filesystem path. Clients can resolve books using content and metadata identifiers, including hashes, before exchanging reading state. Library liseur-sync can index either: a directory containing EPUB files a Calibre library using metadata.db Plain directories derive series information from their directory structure. Calibre libraries use metadata.db ","listed_at":1789162216,"md_updated_at":1789162216},{"full_name":"chmouel/liseur","url":"/r/chmouel/liseur","github":"https://github.com/chmouel/liseur","title":"liseur","tagline":"Open-source EPUB reader and calibre-web client for Android, with a Kindle-inspired reading experience","source":"trawl","tier":"found","license":"MIT","stars":71,"language":"Kotlin","score":0,"critic_up":0,"ai_generated":"mostly","human_touch":"light","status":"works-on-my-machine","category":["mobile"],"built_with":[],"slopbucket":["vibe-coded"],"has_images":false,"vulns":{"deps":0,"vulnerable":0},"pitch":"**The Cap'm wrote this paperwork, not the owner.** This repo never submitted itself to SlopScore. The Cap'm picked it by hand: an open-source EPUB reader and calibre-web client for Android that its owner tagged vibe-coding. It carries the MIT license. The disclosures above are his best guess from what GitHub shows.\n\nIs this yours? Commit a real `slopscore.md` and press Refresh to replace this, or remove the listing in one click. There's no account to make: you log in with GitHub.","readme":"Liseur An open-source ebook reader for Android, and a client for calibre-web, Komga, liseur-sync, Grimmory and any OPDS catalog: EPUBs on your phone, in sync with your own book server. The shelf, sorted by what you are currently reading. Distraction-free page with notes and bookmarks. Themes, open typefaces, spacing, and brightness. More screenshots here About Liseur Liseur renders EPUBs with the Readium engine, in open typefaces (Literata, Vollkorn, Atkinson Hyperlegible, Inter) or whatever the publisher shipped. Four reading themes: Light, Sepia, Dark, OLED Black. Margins, line spacing, brightness, and page-turn vs. continuous-scroll are adjustable per book or library-wide. Auto-scroll runs at a set pace and continues across chapter boundaries. A footer shows time remaining in the chapter. Highlights, margin notes, bookmarks, and dictionary lookups are inline; book-level notes live in the notebook. Marking a passage offers three colours, and you tick which of the six you want along with the colour new marks use. Footnotes open as a card over the page rather than sending you to the back of the book. The library is one shelf whatever the source: local folders, calibre-web, Komga, liseur-sync, Grimmory, or any OPDS catalog. Series are grouped into stacks tracking reading order, progress, and missing volumes. \"Download all books\" fills the shelf from a connected server in one go. Reading position syncs across devices through calibre-web, Komga and liseur-sync, down to the exact","listed_at":1789162212,"md_updated_at":1789162212}]}