tierlist.tech
中EN

09.17 Today's Picks

369 new projects found today, 214 rated.

Created 2026-09-15 – 2026-09-17Fetched 600Filtered 50Rated 34

S · Look today

S

knoxmap

spytheeuclidean-a11yPython

It converts real locations into playable maps for Project Zomboid B42 and explicitly credits upstream Knoxify; a rare genuine need for PZ modders and players who want to fight zombies outside their home, limited by dependence on the upstream toolchain and output quality that varies with geographic data.

KnoxMap: turn any real place into a playable Project Zomboid Build 42 map (built on Knoxify)

#Games#Game Dev#3D#project-zomboid

★34

A · Worth watching

A

obsidian-mobile-companion

zhdmm35Kotlin

Syncs a GitHub Markdown vault to Android, rendering it in Obsidian and handling edit conflicts, filling a gap in the official mobile sync. Useful for self-hosted note users, though it requires a GitHub repo and an extra desktop helper.

Android companion for Markdown vaults stored on GitHub, with Obsidian rendering, offline reading, conflict-safe editing, and an optional desktop sync helper.

#Android#Knowledge Base#Self-Hosted#obsidian

★33

A

claude-vault

KrimchaninTypeScript

Targets the specific pain that switching accounts in Claude Desktop loses sessions, offering local backup and restore that works on install and does only this one thing. 37 stars in one day shows real demand; the limit is covering only the desktop app, so it must follow if the official storage format changes.

Keep Claude Desktop chats safe when switching accounts — local-first backup and restore for conversations.

#Backup#AI#LLMs#Privacy

★37

A

pi-rain-radar

alex-soulJavaScript

A rainfall radar display on Raspberry Pi and Docker with animated playback, offline continuity and multi-screen LAN support. People wanting a dedicated weather screen get a clear deployment path; functionality is single-purpose but complete, and it depends on external radar data sources.

A self-contained Raspberry Pi and Docker rain radar display with animated observed rainfall, local history, outage-resilient playback and LAN multi-screen support.

#Self-Hosted#Hardware#raspberry-pi

★40

A

rimagent

zorrobyteC#

Lets an LLM play RimWorld and rewrite its own strategy manual between runs, with a RimBridge mod, a rare self-improving agent experiment. Interesting to agent researchers and RimWorld players alike, though it requires the game and mod, and results depend heavily on the model.

An LLM that plays RimWorld and rewrites its own playbook between colonies. RimBridge mod + self-improving agent.

#AI Agents#Games#LLMs#rimworld

★26

That's today. The other tiers and the boards follow.

BHas a spark46
  • B typesafe-computer-use Replaces multimodal large models with OCR plus action classification for computer use, at about $0.0002 per step, directly targeting the pain of agents operating computers being too expensive. Its 62 stars in 0.3 days hit a cost-sensitive nerve, and a hosted API is possible; the risk is that native computer use from OpenAI/Anthropic keeps getting cheaper and will crush it head-on. #AI Agents#Automation#macOS

    Computer use for about $0.0002 a step: OCR the screen, classify the next action with TypeSafe, click. macOS.

    awlevinPython★62

  • B bashka Statically verifies script safety before curl | bash and manages installed software in one place, addressing trust in blindly running install scripts. 52 stars in 1.1 days reflect a real pain point hitting supply-chain security. It has potential to become a standard install-step dependency; risk is that static analysis can be bypassed and Homebrew and nix already partly cover the scenario. #CLI#App Security#DevOps#Automation

    Stop running unverified bash scripts to install software! curl <url> | bashKA for static verification of bash script safty and managing installed software

    dmtrKovalenkoRust★52

  • B agent-monitor Visualizes the execution traces of Codex and Claude Code subagents, breaking down tokens and cost. Heavy coding-agent users can use it right away, and it targets a real pain point. The limit is it only covers these two CLIs and must chase them whenever their formats change. #AI Agents#Observability#LLMs#claude-code

    Visualize your Codex and Claude Code subagents. Advanced traces view, tokens breakdown, and cost usage is also supported.

    donvitoTypeScript★21

  • B VFD-GLB-Importer Adds the missing one-click glTF/GLB import to 3ds Max 2024–2026, including PBR materials, hierarchy, skinning, and animation. A clear pain point for those using Max in real-time pipelines, with a specific feature list; limits are covering only three versions and single-person maintenance. #3D#Plugins#Animation

    One-click glTF 2.0 / GLB importer for Autodesk 3ds Max 2024-2026. PBR materials, textures, hierarchy, skin and animation. Created by Fady Derias.

    Derias13Python★8

  • B r-mdir It brings DOS-era Mdir III's two-pane file management and function-key operation back to the terminal in Rust, with a restrained scope and honest description; worth ten minutes for anyone nostalgic for old-style operation or wanting an extremely light TUI file manager, at the cost of a Korean interface and an early ecosystem. #CLI#Terminal#Productivity

    A two-pane file manager that brings the screen, per-extension colors, and function-key-centric operation of the classic Mdir III to the terminal

    rainygirlRust★8

  • B AE_StaffGenerator An After Effects ScriptUI panel that generates staff notation and notes inside a composition, saving manual note placement for score-animation work. Scope is tiny and the description honest; the downside is it only serves a narrow need inside AE. #Animation#Images#Plugins#after-effects

    A ScriptUI panel that generates staff notation and notes in After Effects

    x5arJavaScript★8

  • B QuickPPPoE A portable Windows PPPoE dialer that adds the auto-reconnect the built-in dialer lacks, ready to use for campus and broadband users. Scope is restrained and the description honest; the downside is many similar small tools and limited functionality. #Windows#Networking#Automation

    A lightweight, installation-free Windows PPPoE broadband dialing and auto-reconnect tool

    SakuraChanNyaC#★17

  • B influxdb2tsfile Migrates InfluxDB 1.x/2.x and line protocol to Apache TsFile with schema mapping, parallel export and resumable transfers. Aimed at teams storing long-term time-series data, the problem is concrete and engineering details are solid; the limit is serving only the TsFile ecosystem. #Databases#CLI#IoT

    Command-line tool to migrate InfluxDB 1.x/2.x and line protocol files to Apache TsFile — auto schema mapping, parallel time-sliced export, resumable and verified.

    TimechoLabJava★8

  • B avdslim Cuts Android emulator AVD memory from about 8GB to 1.5GB, addressing the pain of memory-hungry emulators on Apple Silicon and Linux, inspired by simslim. Its 49 stars in 0.7 days show real demand, and it could be absorbed into the Android toolchain; the risk is being squeezed out by official Google optimizations. #Android#Performance#DevOps

    ⚡ Drop Android Virtual Device (AVD) emulator RAM from ~8GB to ~1.5GB on Apple Silicon & Linux. Inspired by simslim.

    kdbhalalaGo★49

  • B Switch-PC-Stream Uses an unmodded Switch Lite as a 720p60 streaming secondary screen; the idea is novel and the scope small. Worth ten minutes for anyone with a Switch who wants a second screen. The limit is reliance on a specific streaming chain, with no measured latency or image-quality data yet. #Games#Streaming#Windows#nintendo-switch

    Low-Latency 720p 60 FPS PC Game Streaming on a Stock, Unmodded Nintendo Switch Lite

    RaZOr519Python★11

  • B vscode-mcp-api Lets web-based AI call local MCP tools through a VS Code extension to read and write files, targeting the real gap that ChatGPT's web version cannot reach the local environment. The idea is direct and small in scope, but it hands local file permissions to web AI, so you must weigh the security boundary yourself. #MCP#Plugins#Editors#AI

    Let web AI easily use MCP tools to manage local files

    DanchunaTypeScript★10

  • B CustomDance A SIGGRAPH Asia 2026 paper repo on 3D dance generation using coarse-to-fine human interaction control, a novel direction suited to animation and embodied AI researchers. The limit is the paper code isn't yet proven reproducible, so ordinary developers can't use it. #ML Research#3D#Animation

    [SIGGRAPH Asia 2026] CustomDance: Customized 3D Dance Generation with Coarse-to-Fine Human-Centered Interactive Control

    XulongTPython★28

  • B easy-unlocker A tool for managing AI agent secrets: biometric approval on your phone, end-to-end encryption straight into the target process, with plaintext never touching chat or logs. Teams deploying agents will care. The catch is it needs a companion phone app and the ecosystem is still tiny. #Secrets#AI Agents#Privacy#Android

    Secrets management for AI agents: approve API keys and credentials on your phone with biometrics, delivered end-to-end encrypted into the target process. Plaintext never enters the chat, logs, or rela…

    cyancityKotlin★9

  • B awesome-cloudflare-selfhosted Curates open-source apps that can replace SaaS inside your own Cloudflare account, solving product discovery; 265 stars in 1.6 days shows the Cloudflare self-hosting topic is hot. The path is becoming the standard index for this niche, sponsorship, or cloud-vendor deals, but awesome lists are heavily homogenized and easily forked, capping commercial upside. #Self-Hosted#Knowledge Base#DevOps

    🍊☁️ Open-source* apps that replace a SaaS product, running in your own Cloudflare account

    theoephraimJavaScript★265

  • B privacy-router Privacy routing for local AI requests: detect sensitive content, confirm once manually, then automatically block or allow based on the registered decision. The problem is well chosen, since local models also need data boundaries, but currently there is only a concept-level description and 8 stars, so whether it lands remains to be seen. #Privacy#AI#LLMs#Self-Hosted

    Local AI privacy routing: discover sensitive content, review once, enforce registered decisions on future requests.

    Inling-AIRust★8

  • B qr A pure-browser QR code generator that exports SVG/PNG, with minimal code and no extra features, showing good restraint from the author; worth a look if you want an offline local QR tool, though similar tools are everywhere. #Productivity#Web Frontend#Images

    A simple, browser-only QR code generator with SVG and PNG export

    mitsuhikoJavaScript★8

  • B AHCINT A universal SATA AHCI driver for Windows NT3.5 through XP, letting old systems recognize new drives. Retro PC enthusiasts can use it immediately and the work is substantial; the limit is a very niche platform and non-trivial installation. #Windows#Hardware#Embedded

    Standard SATA AHCI Controller Driver for Windows NT3.50, NT3.51, NT4.0, 2000 and XP x86 and x64

    ages2001C★8

  • B ResetSpy It enumerates tenant accounts and registered authentication methods through Microsoft's self-service password reset portal, a common information-gathering technique in red teaming and audits, packaged as a standalone script for convenience; useful for identity security assessments, but it could also be used to probe accounts. #Network Security#App Security#microsoft

    Enumerate user accounts and registered authentication methods via the Microsoft Self-Service Password Reset (SSPR) portal

    mlcsecPython★18

  • B closetscan Turns a walk-around-your-closet video into a searchable clothing catalog, with a local MCP server and agent skill, targeting a specific and rare scenario. But it depends on capture quality and a multi-model pipeline, and at this scale stability isn't proven. #MCP#Multimodal#AI Agents#Video

    Turn a closet walkthrough video into a searchable wardrobe catalogue, with a local MCP server and agent skill

    wgaostudioPython★7

  • B ninfer-v100 Deployment scripts that get models like Qwen3 running on V100 (sm_70), with systemd services and a comparison against llama.cpp. Anyone with a V100 has been missing this path. The limit is that it's a fork of someone else's project with narrow coverage. #LLMs#GPU#Performance#Benchmarks

    NInfer on Tesla V100 (sm_70): Qwen3.8-27B deployment, build/convert scripts, systemd service, and cross-engine benchmark (vs llama.cpp). Fork of Neroued/ninfer with sm70 vision-encoding speedup.

    dollarwongC++★15

  • B scan2mesh Aligns lidar scans with a prebuilt 3D mesh for localization and tracking; the goal is focused and the implementation is C++, making it real engineering. But it doesn't specify the middleware or data formats it depends on, so onboarding cost is unknown; those doing robot localization may want a look. #Robotics#3D#Performance

    Scan2Mesh is a high-performance mesh-based LiDAR localization and tracking system that estimates LiDAR poses by aligning incoming scans with a pre-built 3D mesh.

    fuxingyinC++★10

  • B open-transformers Open-sources a model family's pretraining pipeline together with audit replay tools, emphasizing reproducible training and verifiable process, still rare in public pretraining. Useful for teams doing pretraining and compliance audits, though the bar is high and documentation quality remains to be verified. #LLMs#ML Research#Benchmarks#GPU

    Reproducible LLM pretraining and audit-replay harness behind the OPEN model collection

    gensyn-aiPython★10

  • B AirPoise Uses AirPods motion sensors for posture reminders and head-gesture shortcuts, running entirely on-device without a camera; repurposing existing hardware is a clean idea worth ten minutes for Mac users. The limit is it only covers the AirPods ecosystem. #macOS#Health#Accessibility#airpods

    Mac menu-bar app: AirPods posture coaching and head-gesture shortcuts. On-device, no camera, no cloud.

    jaskirat1616Swift★10

  • B car-systemui-win98-pod It turns the Windows 98 window look into a system UI plugin for Android Automotive, with draggable, resizable windows that take over real app tasks; a rare little toy for car ROM tinkerers, but it requires an Automotive OS 17 environment, so almost no one can verify it firsthand. #Android#UI Design#Windows

    Windows 98 window chrome for Android Automotive OS 17, built on Scalable UI. Drag, resize, minimize and maximize real app tasks. Ships as a CarSystemUI pod.

    passenger6Java★8

  • B EasyThumbnailGenerator A UE5 editor plugin that renders static and skeletal meshes into transparent PNG thumbnails with presets and batch mode. Artists and programmers building asset libraries and store images can use it directly; scope is restrained, though similar plugins already exist. #Game Dev#Plugins#3D#unreal-engine

    Unreal Engine 5.8 editor plugin for generating transparent PNG thumbnails from Static Mesh and Skeletal Mesh assets. Includes live preview, camera and lighting controls, reusable presets, batch genera…

    ZUPTIC++★7

  • B livediff A small command-line tool that shows git diff in real time, with its purpose stated in one sentence and no extra ambition; but lazygit, tig plus delta already cover most scenarios, so there is limited novelty. #CLI#Code Quality#Terminal

    A live Git diff view cli

    stagasGo★7

  • B dev101x-pentest-lab A Docker-packaged pentest practice range chaining blind SQLi, hashcat, path traversal, command injection, internal pivoting, and SUID privilege escalation, with flag validation. Concrete and ready to run, but similar ranges like Juice Shop, VulnHub, and HTB are already crowded. #App Security#Network Security#Exploits#Testing

    Self-contained Docker Pro Lab for pentest practice: boolean blind SQLi + hashcat cracking, web path traversal, command injection, internal network pivoting and SUID PATH-hijack privesc — with live fla…

    DevCop95Python★8

  • B a-rhythm-trainer A browser-based rhythm trainer for tapping practice and rhythm dictation, with preset, random, and custom questions. The features are clear and the scope restrained, suitable for music learners; but rhythm-training web tools are already plentiful, so novelty is limited. #Audio#Learning#Web Frontend

    A rhythm training tool supporting tapping practice, rhythm dictation, and preset, random, and custom exercises

    Emmet-RayTypeScript★14

  • B kakao-headless Turns KakaoTalk into a local CLI with encrypted credentials and send confirmation, letting scripts read messages and batch-send. A rare automation entry point for Korean developers; but it relies on an unofficial client protocol, so account risk and terms-of-service issues are on you. #CLI#Social#Secrets#Automation

    Read. Approve. Send. A local KakaoTalk CLI with a managed Aside integration, guarded sends and encrypted credentials.

    JSap0914JavaScript★10

  • B BareDroid Starts Linux natively on Android via Termux, bypassing chroot. Few have tried this direction, but it is self-labeled experimental and likely works only on certain models, so ordinary developers cannot install it today. More a tech demo; wait for specific models and images. #Linux#Android#Operating Systems

    [EXPERIMENTAL] Full Linux environment on Android phones — native boot, no chroot, Termux, PRoot, containers, or Android compatibility layers.

    RubencskuPython★7

  • B Moss-Transcribe-WebUI A local transcription WebUI built on transcribe.cpp with speaker diarization, and the localization angle holds up. But Whisper-based WebUIs are already crowded, and it doesn't clarify speed or accuracy differences, so watch for follow-up. #Speech#Multimodal#Self-Hosted

    A local high-performance speech transcription WebUI based on transcribe.cpp + Moss-Transcribe-Diarize.

    TigerSHe1998Python★7

  • B resolvr-dns-dashboard A self-hosted monitoring panel for Technitium DNS with real-time query logs, cache and blocking stats, plus login and roles. Technitium users would want it, but the server's built-in interface already covers most of these features. #Monitoring#Self-Hosted#Networking

    Self-hosted monitoring dashboard for Technitium DNS Server — live query logs, clients, cache, and blocking, with its own login and role-based access.

    bassem-elsodanyTypeScript★13

  • B seenix Draws the Nix store closure into the browser at one byte per pixel, with zoom and pan, to show visually what's eating disk space; the idea is fun and the implementation restrained. But only Nix users can use it, making it a playful diagnostic rather than a necessity. #Linux#Web Frontend#Performance

    Pan around a Nix closure in your browser, one byte per pixel

    fzakariaJavaScript★13

  • B xcode-project-format Claims to read and write Xcode's JSON project format, but Xcode projects are still pbxproj, so the description doesn't match reality, and being under the apple org makes source verification more necessary. Not recommended to rely on until clarified. #macOS#Build#xcode

    A Swift library for reading, writing, and manipulating Xcode's JSON-based project.xcproj format.

    appleSwift★25

  • B qwen38-exl3-dflash2 Configuration notes for quantizing Qwen to 4bpw and running 262k context on a 24G 3090, with specific parameters that give reference values to people running local models on the same GPU; but it reads more like personal experiment notes and does not apply to other hardware. #LLMs#GPU#Performance

    Qwen3.8-27B EXL3 (4.00 bpw) + DFlash2 speculative decoding for ExLlamaV3, validated at 262k context on a 24 GB RTX 3090

    r0b0tlabPython★9

  • B ComfyUI-FS_Audio_Suite Breaks YuE2 music generation into ComfyUI nodes for easy integration into existing audio workflows. Useful for ComfyUI music users, but node packs are already crowded and YuE2 itself is early, so its interface may change at any time. #Audio#AI#Plugins

    FS_Audio Suite: modular YuE2 audio generation nodes for ComfyUI, by Make the Robot Do It

    KytraScriptPython★25

  • B frappe_config_pack A portable, versioned configuration package for the Frappe framework that addresses configuration drift during deployment, a pragmatic direction. The audience is limited to self-hosted Frappe teams, value is narrow, and full usage docs are still missing. #DevOps#Self-Hosted#frappe

    Portable, safe, versioned configuration packages for Frappe Framework.

    abdo-hostPython★8

  • B owasp-agentic-lab A weekend experiment designed around the OWASP Agentic Top 10: build an agent, then break it five ways and harden it. It has teaching value for agent security practitioners, but it's essentially course material with no reusable tool left behind. #AI Agents#App Security#Learning

    Weekend lab for the OWASP Top 10 for Agentic Applications. Build an agent, break it five ways, then harden it.

    bytemonk-academyTypeScript★7

  • B quietstart A control-recognition tool for HarmonyOS native pages that uses prompt learning to generate rules automatically, filling a gap in HarmonyOS automation; but the description is vague and lacks usage and accuracy data, making it hard to judge whether it works reliably. #Automation#Testing#harmonyos

    A HarmonyOS native page control recognition assistant tool. Recognize and learn pages following prompts, and automatically generate rules.

    thx853068675-devLanguage unknown★25

  • B shanhui-ai-kefu An AI customer-service desktop app covering six platforms including Taobao and Douyin Shop, learning in-store product specs before answering and handing off to humans when unsure, a real need. But there's only a description, no code or deployment docs, and multi-platform adaptation costs raise doubts about maturity. #AI#E-commerce#Automation#Windows

    E-commerce AI customer service bot / AI customer service system: unified reception across six platforms and multiple stores — Qianniu (Taobao/Tmall), Xianyu, Douyin Store (Feige), JD (Jingmai), Kuaish…

    wss7750Language unknown★10

  • B worky.nvim A git worktree switcher for Neovim, with a one-line description and clear boundaries. The problem is that worktree plugins and neogit already offer similar capabilities, and it covers only a tiny piece, so it is not worth installing separately unless you happen to lack this keybinding. #Plugins#Editors#Productivity

    A git worktree switcher for neovim

    GRVYDEVLua★10

  • B hotseat Shows quota usage for multiple Claude Code and Codex accounts in the terminal, then continues work interrupted by rate limits. The need is real and the scope restrained; but usage-panel tools have clustered recently, and 6 stars is not enough to prove it is more stable. #CLI#Terminal#AI#Monitoring

    Terminal dashboard and CLI for Claude Code and Codex accounts, their quota, and work a usage limit stopped

    MaxteabagGo★6

  • B confset It uses Pkl to generate local config files for various tools, consolidating configuration into one verifiable source, with a small scope and restrained naming; but templates plus scripts already suffice, and it does not say which tools are supported, so the advantage is unclear. #Build#CLI#DevOps

    Generate native tool configuration from Pkl

    mathematic-incRust★8

  • B glimmer A cross-platform Chinese input method written in Rust, explicitly stating it derives from Qingjian and follows GPL, which is honest; input method engines have a high barrier, and this is only an early derivative, far from replacing RIME day to day. #Productivity#Linux#Windows

    Weiming Glimmer: a cross-platform input method written in Rust, derived from Qingjian (GPL-3.0-or-later)

    rambocodeRust★9

  • B uptime-pulse A lightweight self-hosted uptime monitor written purely in Rust, with a clear stack and low memory as its selling point. But Uptime Kuma and Gatus already cover this space thoroughly; unless you care deeply about resource use, there's no reason to migrate. #Monitoring#Self-Hosted#DevOps#Observability

    Ultra-lightweight self-hosted uptime monitoring system in pure Rust (Axum, Tokio, SQLite WAL, Rustls). Fast, low memory, and multi-arch ARM64/AMD64 ready.

    zidnyzdRust★7

  • B shiyi-ai-diet-assistant A live WeChat mini-program that uses multimodal recognition to log meals with nutritional guardrails, a more complete product idea than typical demos. But the diet-tracking space is crowded, and the repo doesn't clarify what's open-sourced or how to self-deploy. #Multimodal#Health#AI#wechat

    Shiyi: a launched AI diet-logging WeChat Mini Program using multimodal recognition, Human-in-the-loop, and deterministic nutrition guardrails

    zj83339-debugJavaScript★7

COrdinary70
  • C VpsCT A self-hosted VPS management panel for individuals and small teams, centralizing status, traffic, service configuration and resource sharing. Its 91 stars in 1.5 days show real demand for lightweight ops panels, with monetization via subscription or hosting; but Baota and 1Panel already hold the space firmly, and Go panels have low feature barriers, making differentiation hard. #Self-Hosted#DevOps#Monitoring

    A self-hosted server management panel for individuals and small teams, centrally managing VPS status, traffic, service configuration, and resource sharing.

    YongshengWinGo★91

  • C github-stars-history Tracks and visualizes star growth curves for any repository, serving project growth analysis and selection decisions. Its 227 stars in 0.3 days come from developers' natural curiosity about growth data, but star-history.com and similar tools already occupy the space, the feature can be easily copied by official Insights or third parties, and there is little reason to pay. #Data Analysis#DevOps

    Track and visualize the stars history of any GitHub repository. Open-source growth analytics and velocity tracking.

    Marcos66236Python★227

  • C xiaohongshu-api Provides Python call examples for searching Xiaohongshu notes, details and comments, serving content analysis and marketing data needs. Its 45 stars in 0.3 days come from scraping demand, and selling a data API is possible; but the platform keeps upgrading anti-scraping measures, legal and compliance risks are high, and competitors are many mature data providers. #APIs#Social#Data Analysis

    Xiaohongshu API, Xiaohongshu interface: Python call examples for note search, note details, and comment queries. Just One API.

    justoneapi-labsPython★45

  • C expo-gpt-live A mobile voice app scaffold based on Expo SDK 57, integrating GPT-Live and AI Elements personas to help developers quickly build real-time voice agents. Its 72 stars in 1.2 days come from topic hype, but it is essentially a demo/starter that official templates could replace, lacking a sustainable productization path. #Speech#AI#LLMs#Multimodal

    Expo SDK 57 voice app with GPT-Live 1 and AI Elements Persona.

    davidmokosTypeScript★72

  • C cve-2026-41940-PoC An exploit tool for a cPanel/WHM authentication bypass that lets attackers skip login to the control panel. Its 72 stars in 0.4 days show broad impact and quick attention from the security community; there is no commercialization possibility, but it is worth defenders' attention as an in-the-wild threat signal, with the risk that PoC spread accelerates real attacks and vendor patch speed is key. #Exploits#App Security#Network Security

    A cPanel and WHM authentication bypassing tool

    ctdalPython★72

  • C rg-sp-ds-style Turns the stock system of the RG SP GBA handheld into an NDS-style launcher, reusing existing emulators and saves. The audience is a tiny group of handheld players, but the author states the applicable scope honestly, making it a clean small modification. #Games#UI Design#Linux

    DS Style for RG SP: a Nintendo DS-inspired GBA launcher ported to stock OS, using your existing emulators and saves.

    FrankieT19C★8

  • C sc64-savestates Adds save states, a virtual controller memory card and slow motion to the SummerCart64 flashcart, as a fork of N64FlashcartMenu; the changes are concrete and aimed at real hardware players. But the audience is tiny and it exists only as a patch. #Games#Embedded#Hardware

    Save states, a virtual Controller Pak, suspend/resume and slow motion for the SummerCart64: a fork of N64FlashcartMenu, one file to drop on the card

    john-steinmeyerC★10

  • C SoDam_O-Brain-Codex It stores decisions and preferences from Codex sessions in a local SQLite database, retrieves them via an MCP plugin, and renders them as graphs and timelines; the idea is practical, but memory-layer projects are already crowded, and this setup is not hard to build yourself. #MCP#AI Agents#Knowledge Base#LLMs

    A local-first MCP plugin that remembers the decisions, constraints, and preferences of Codex conversations in local SQLite and manages them with search, 2D/3D graphs, timelines, and backups

    sodam-aiHTML★8

  • C omarchy-key-promoter Brings Key Promoter X to the Omarchy desktop, showing themed hints when users perform actions with the mouse that have shortcuts, helping people memorize keybindings. Its 71 stars in 1.2 days come from the desktop-customization community, but the platform is very narrow, the feature is single-purpose, there is no monetization path, and competitors are built-in keybinding hints in each IDE. #Linux#Productivity#Plugins#omarchy

    Key Promoter X for your Omarchy desktop: themed toasts that show the keybinding for what you just did the slow way

    ForrestKnightQML★71

  • C ai-coding-interview-practice Ten problems with pre-planted bugs plus timers and interviewer hints, for practicing AI-assisted coding interviews. The format is somewhat interesting, but it's ultimately quiz content with limited value for readers seeking tools. #Learning#AI Agents#Jobs

    Practice the AI-assisted coding interview format with your own coding agent: ten problems with planted bugs, stubs, timed tests, and an interviewer prompt

    azizu06Python★15

  • C ipad_computer_use Lets agents operate an iPad through USB input tools and native screen mirroring and control services, expanding computer use's device coverage. Its 49 stars in 1.3 days come from agent hype, but it relies on private protocols, the platform is closed and easily broken by system updates, productization is hard, and competitors are mature mobile automation frameworks like Appium. #Automation#iOS

    Control an iPad through a USB input tool, native screen sharing, and a control server.

    jamiepinheiroJavaScript★49

  • C ruby_llm-typesafe Connects a vendor's structured output into RubyLLM; the code should be small and the naming honest. But it's one provider added to one gem, meaningless outside Ruby LLM users, and the README still lacks usage examples. #LLMs#SDKs#APIs

    TypeSafe structured-output provider for RubyLLM 2

    kieranklaassenRuby★9

  • C product-motion-gif It has a model generate multi-panel frames in one pass and stitches them into a looping GIF, aimed at people who cannot use design software to make product animations, with clear inputs and outputs; but the cited model names do not match reality, and similar image-to-animation tools are already plentiful. #Image Generation#Agent Skills#Animation#AI

    Who says AI can't generate GIFs? This skill uses GPT-Image 2.5 to generate multi-panel frame images in one go - composite them into an animated GIF - and produce a dynamic image in a single call. Prod…

    liangdabiaoPython★17

  • C pureterm An SSH/SFTP client built on xterm.js and ssh2, with the tech stack clearly stated; but the terminal client space already has mature alternatives like Tabby and Electerm, and this offers no reason to choose it instead. #Terminal#Networking#CLI

    PureTerm is an SSH/SFTP client built with Cordis, ssh2, and xterm.js

    puremixaiTypeScript★14

  • C busters-hidden-treasure-pc A native PC port of the MD game Tiny Toon with widescreen, instant save, and an in-game menu; you must supply your own ROM. Attractive to retro players and the author honestly states the prerequisite, but fan ports are a mature category with a very narrow audience. #Game Porting#Reverse Engineering#Games#Windows

    A native PC port of Tiny Toon Adventures: Buster's Hidden Treasure (Sega Mega Drive / Genesis). Widescreen, save states, filters, and an in-game menu. Bring your own US ROM.

    DarhkfoxLanguage unknown★7

  • C LivelyAirRaidAlert Turns Ukrainian air-raid alert data into a dynamic wallpaper usable with Lively Wallpaper, of real value to local users. The scope is very small and depends on Windows and a third-party data source, so it means little outside this scenario. #Windows#Monitoring#Web Frontend

    A custom adaptation of data and visual materials from alerts.in.ua and neptun.in.ua designed specifically for use as an interactive web-wallpaper in Lively Wallpaper.

    FullyRealistHTML★8

  • C zhimu-player A Windows player that combines subtitles, timestamped notes, and ChatGPT/DeepSeek/Tingwu transcription into a bilingual video reading workbench. The idea is targeted, but connecting three services and multiple input modes at once makes the scope large, and actual experience remains to be verified. #Video#AI#Speech#Windows

    Zhimu Player — Windows video player with subtitles and timestamped AI guides. ChatGPT web, DeepSeek API, Tingwu ASR. A bilingual Chinese-English video reading workbench.

    OJY-lawyerTypeScript★10

  • C dspy-typesafeify Adds a decorator to dspy Signature that automatically wires in TypeSafe for structured-output constraints; 49 stars in 1.2 days rides the DSPy and structured-output hype. But it is a small upstream-ecosystem plugin hard-dependent on both dspy and TypeSafe APIs, easily replaced, and hard to commercialize independently. #LLMs#SDKs#Code Quality

    Add a decorator for dspy Signatures that automatically uses TypeSafe where relevant

    typesafeainatePython★49

  • C liquid-glass-react A React component library recreating Apple Liquid Glass with edge refraction and frosted-glass fallback, and the implementation looks serious. But such recreations are already crowded, and with only six stars and no demo comparison it's hard to justify replacing existing options. #Web Frontend#UI Design#react

    React component library recreating the Apple Liquid Glass look: real edge refraction, dynamic highlights, and an elegant frosted-glass fallback. | React component library recreating Apple Liquid Glass…

    TsdsjTypeScript★6

  • C FingerChord.app Adds macOS trackpad gestures to close windows and quit apps, with haptic feedback; the idea is practical. But the description's macOS 27 and "generated in one shot by GPT 6 Astra" don't match reality, denting credibility, so whether it runs needs verifying first. #macOS#Productivity#Accessibility

    Native trackpad gestures for macOS 27: ⌘ click, close windows, and quit apps. Haptic feedback, no Dock or menu bar icon. Built one-shot with GPT 6 Astra.

    heruoxinSwift★11

  • C Pie Provides a native macOS graphical interface for the Pi Coding Agent, making the command-line coding agent easier to use. Its 47 stars in 1.6 days come from coding-agent hype, but it is deeply tied to a single upstream project and shares its fate, with barriers only in the UI and competitors being official IDE plugins from various agent vendors. #macOS#AI Agents#CLI#Editors

    Native macOS GUI for Pi Coding Agent

    hieunc229Swift★47

  • C telegram-clone-worker Builds a Telegram channel cloning and real-time sync engine on Cloudflare Workers and D1, extendable into content backup and migration services. With 46 stars in 0.8 days, it touches platform ToS and copyright boundaries, competes with many existing TG download bots, and has limited commercial space. #Social#Downloaders

    Telegram Clone Worker - Fast Telegram channel cloner & live sync engine on Cloudflare Workers + D1

    iamLiquidXTypeScript★46

  • C xiaomi-performance-tray A tray toggle for performance mode on Xiaomi laptops that can link to Windows power-saving settings; the author honestly states it was verified only on the REDMI Book Pro 16 2026. Honesty earns credit, but the audience is limited to owners of the same model. #Windows#Hardware#Productivity

    Xiaomi laptop performance mode tray tool, double-click to toggle and link with Windows Energy Saver. Verified only on REDMI Book Pro 16 2026 (X7 358H), feedback welcome. | Xiaomi laptop performance tr…

    OJY-lawyerC#★8

  • C FlyDrones Uses the fruit fly connectome as a drone flight controller, with camera input passing through compound eyes and a spiking neural network to descending neurons, viewable in real-time 3D in the browser. It is highly novel and scientifically fun, but very far from engineering practicality and commercialization, more likely a research demo or paper companion that developers cannot depend on. #Robotics#Biotech#ML Research#GPU

    A fruit fly connectome as a drone pilot. Camera -> fly eyes -> spiking brain -> descending neurons -> drone. Live 3D demo in your browser.

    SpikeCallsPython★48

  • C KickPersian-Font Changes only the Persian font in Kick chat while leaving English untouched, built on MV3 with a clear goal. The audience is very narrow but the need is real; font-replacement extensions are already plentiful and it offers no differentiator beyond a screenshot. #Browser Extensions#Streaming#UI Design

    A Chrome Extension (Manifest V3) that changes Persian text on Kick (kick.com) to your chosen font, with special focus on streamer chat (including OBS popout). English text stays untouched.

    TheGreatAziziJavaScript★8

  • C validador-transferencias-pymes Real-time transfer verification for Paraguayan small merchants, claiming zero-knowledge fraud prevention; the problem is specific. But the description piles on jargon with no deployment method or interface details, so actual usability can't be judged. #Fintech#APIs#Privacy

    Sistema open-source en TypeScript de validacion de transferencias bancarias en tiempo real para cajas y PYMEs (Itau / SIPAP Paraguay) con Zero-Knowledge y proteccion antifraude.

    fugc123TypeScript★14

  • C WeChatAlive Hooks an OpenAI-compatible group chat bot onto Windows WeChat — not a new idea, already covered by projects like chatgpt-on-wechat, and it relies on client automation, leaving stability and ban risk entirely to the user. The description is clear, but its value depends on whether you'll take that risk. #AI#LLMs#Social#Windows

    Bring your Windows WeChat account to life: an AI group chat bot supporting any OpenAI-compatible model

    Aiden-723Python★9

  • C Metal-Gear-Solid-VR-Standalone A fan-made Metal Gear VR standalone build for Quest 3, still in beta, with 31 stars showing fellow fans are interested. But copyright inherently limits it and completeness is unknown, so ordinary developers can only watch, not use it. #Games#Game Porting#3D#Android

    Versión Beta en desarrollo MGS VR Standalone para Quest 3 hecha por un fan (Fan-made standalone MGS VR beta version for Quest 3 currently in development.)

    BaywillyLanguage unknown★31

  • C RaceWave46 Claims to use AI-assisted recompilation of Wave Race 64 with modern graphics and local multiplayer. N64 recompilation is this year's hot topic, but the repo is only hours old with no build artifacts or instructions, so treat it as another attempt to ride the hype. #Game Porting#Reverse Engineering#Games#Windows

    An AI-assisted Wave Race 64 recompilation for Windows, with modern graphics and expanded local multiplayer.

    DomazinUSC++★17

  • C yanbian A Word assistant for Chinese writing and revision that connects to GPT/DeepSeek, with a license limited to personal non-commercial use and attribution required for lawyers' practice. The scenario is specific, but similar Word AI plugins are many, and the license terms will deter many potential users. #Documents#AI#LLMs#Windows

    Yanbian: a Windows Word writing and revision assistant supporting GPT (Codex) and DeepSeek. For personal non-commercial use only, except for lawyers in practice, and attribution must be retained.

    OJY-lawyerJavaScript★9

  • C resume-filler-extension A browser extension that fills job application resumes with one click, a real need but easy to build with many competitors. The repo doesn't say which sites it supports or whether resume data stays local or is uploaded, a first hurdle for privacy-conscious job seekers. #Browser Extensions#Productivity#Jobs

    A plugin for quickly filling in online job application resume information

    Suara17JavaScript★9

  • C omarchy-green-mosaic-theme A light theme for the Omarchy desktop pairing beige mosaic with dark green, install-and-go and restrained in scope; but color themes have too many alternatives, it's nearly useless outside Omarchy users, and with only a dozen or so stars it isn't worth a special look. #UI Design#Linux#Productivity

    A light Omarchy theme with warm beige mosaic tiles and a matching forest-green palette.

    ejuroCSS★13

  • C Practical-Network-Security-Lab Thirteen modules of network attack-and-defense labs, from reconnaissance and man-in-the-middle to DNS/DHCP attacks and defenses, clearly orchestrated. But it's essentially a teaching lab manual, and with plenty of similar ranges and courses already, innovation is limited. #Network Security#Learning#App Security

    Practical network security lab — 13 modules covering recon, MITM, DNS attacks, DHCP, router exploitation, cryptography, and defense. Isolated VM lab. Attack → Detect → Fix.

    imranshaikh116Python★10

  • C openjarvisbr A Portuguese-language voice assistant written in Rust, wired to Gemini Live for real-time conversation. Voice assistant wrappers are everywhere, and this one is half a day old with no install steps or latency data; twenty-odd stars look like early gawking, not usage. #Speech#LLMs#AI Agents

    OpenJarvisBR — assistente de voz open source em Rust com Gemini Live

    ustoppbleRust★26

  • C pi-jev Adds a typed-validation gate to tool calls for the Pi coding agent, motivated by constraining agents from calling tools wildly. But it never explains what "jev" is and offers no comparison data, so readers can't judge whether to replace their current setup. #AI Agents#LLMs#APIs

    TypeSafe Jev as a decision layer for the Pi coding agent: a measured tool-call gate plus jev_ask for typed, calibrated answers

    y0usafTypeScript★8

  • C hermes-memory-wiki Adds a session-history wiki and a long-term memory audit panel to Hermes Agent, laying out what the agent remembers for people to see. The direction is interesting, but it depends on a not widely known agent, and 7 stars show no user base has formed yet. #AI Agents#Knowledge Base#Plugins#LLMs

    Memory Wiki dashboard plugin for Hermes Agent: browsable session-history wiki + persistent-memory audit panel

    NousResearchPython★7

  • C WCAE A platform that collects WeChat public account articles and claims to connect to Agents. Collection-plus-Q&A approaches already exist in many forms, and the repo doesn't clarify retrieval methods or data compliance, so practical value remains unverified. #AI Agents#Automation#Content Writing

    A WeChat Official Account article collection platform with Agent integration support

    YMACPC#★11

  • C typesafe-mario Drives an agent through Super Mario Bros. using structured emulator state, separating the perception interface from model decisions; it's a clean little experiment. But with no benchmarks or reuse notes, it's closer to a personal demo than a tool. #AI Agents#Games#LLMs

    A TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state.

    fhshaikPython★12

  • C hermes-filebox A file explorer for Hermes Desktop with tabs, previews and 3D viewing; it packs in plenty of features. But it's tied to a niche desktop shell, so outside users can neither install it nor need it. #Productivity#Plugins#3D

    Explorer-style file browser for Hermes Desktop: tabs, previews, edit ops, 3D viewer

    imzacksongJavaScript★22

  • C meow-memorycarry-zcode A session-memory handoff plugin for ZCode: /handoff saves a Markdown memory store and /recall reloads it in a new window. The idea is practical but tied to an obscure tool, with under 10 stars, so outside readers basically cannot use it. #AI Agents#Agent Skills#Knowledge Base#Plugins

    ZCode session memory handoff plugin: /handoff stores a window's knowledge into a layered Markdown memory library, /recall reloads it in a new window.

    LogocceaiJavaScript★10

  • C folio A launcher that gives foldable Android phones an iPhone-style home screen, without root, plus a few jailbreak-flavored tweaks. The aesthetic is clear, but 'iOS-style launchers' are nothing new in app stores or the open-source community. #Android#UI Design#Productivity

    Folio Launcher: a clean, iPhone-style (iOS) Home Screen for Android foldables like the Galaxy Z Fold, with jailbreak-inspired tweaks. Free, open source, no root.

    McCal-CodesKotlin★14

  • C Aria Renders osu! storyboards as dynamic desktop wallpapers, implemented in Rust. An interesting little tool for osu players, but the audience is very narrow and you must prepare your own beatmaps and assets, so outsiders have no use case. #Animation#Games#Audio#Windows

    Aria - osu! storyboard dynamic desktop wallpaper

    OhdmireRust★8

  • C yappynotes A cross-platform sticky notes app in Avalonia with MVVM and SQLite, a solid technology choice. But note apps are oversupplied and it states no differentiator anywhere, making it hard to switch from the one you already use. #Productivity#macOS#Windows

    Cross-platform desktop sticky notes - C#/.NET 10, Avalonia UI, SQLite. MVVM with a repository.

    algorisys-ossC#★8

  • C uwlearn_assignment_extension Makes Waterloo Learn assignment lists and deadlines clearer and syncs them to a calendar, a small delight for Waterloo students. But it serves a single campus system, so scale and generality are limited and it isn't a project worth watching. #Productivity#Browser Extensions#Learning

    Make viewing assignments and due dates easier on Waterloo learn + integration into calendar.

    gurshh-rainJavaScript★9

  • C AnyWorld A browser multiplayer game where an AI is the dungeon master and players act in natural language. Letting an LLM be DM has been tried repeatedly these past two years, and this doesn't say how rule adjudication, latency or multiplayer sync are handled, so no new approach is visible. #Games#LLMs#AI Agents

    A multiplayer browser game implemented with Python, where an AI acts as the dungeon master and players type out their actions in their own words!

    iamarxsPython★8

  • C VanillaPlusMegapack A Lua mod that bundles nine Helldivers 2 improvements, with enemy-faction prediction before deployment as its selling point. Useful for mod users, but it is only a mod collection, not a new tool or method, so outsiders have no use for it. #Games#Game Dev#Plugins

    Nine Helldivers 2 improvements in one package, including Know Your Constellation for enemy forecasts before deployment.

    CowboyBingusLua★13

  • C CLIRPG A pure terminal roguelike with a rune system and endless rift setting; the name and scope are clear. But terminal roguelikes already have many mature works, so this is just another personal mini-game, not a new approach. #Games#Terminal#CLI

    Endless Rift — a pure-terminal roguelike. Descend an endless rift, hoard loot, master the rune system, ascend the Rift.

    DapoletC++★7

  • C DeltaHarmonicaPlayer Plays harmonica via MIDI in Delta Force, with numbered-notation conversion and offline transcription. The feature is specific and somewhat interesting, but the audience is a tiny slice of players, in-game performance risks a ban, and it has no general value. #Games#Audio#Windows#delta-force

    Delta Force PC game harmonica MIDI playback, numbered musical notation conversion, and offline AI audio transcription tool

    hcst06C#★8

  • C east-china-sea-red-tide Packages historical East China Sea red tide queries and LightGBM next-month prediction as a 'skill,' complete with a judge demo flow. Both data and model come from a specific project, so generality is limited; it looks more like a competition entry than a reusable tool. #Science#Data Analysis#ML Research

    East China Sea red tide risk assessment skill: historical queries, real LightGBM next-month forecasts, reproducible programs, and a judge demo

    Hu-yooukPython★12

  • C JourniOne-Planning-Skills An agent skill that breaks travel planning into inspiration, day-by-day routes, and a shareable Travel Journal, and also wants to link hotels and flights. It sounds complete but does several things at once and carries obvious product promotion, so the real trade-offs are not yet visible. #AI Agents#Agent Skills#Productivity

    JourniOne travel planning Skill: from inspiration and day-by-day itineraries to a shareable Travel Journal with maps, connecting hotels and flights as needed.

    JourniOne-aiJavaScript★13

  • C herdr-agent-progress Adds a self-reported task progress line to the Herdr sidebar; the idea is straightforward and the implementation should be small. But it's locked into the tiny Herdr ecosystem, and the description never explains integration or actual results, so outsiders can't judge its value. #AI Agents#Plugins#Observability

    Agent-reported task progress and activity for the Herdr sidebar

    eliasstravikRust★14

  • C yingxian-pagoda Tells the story of the Yingxian Wooden Pagoda through a six-act scrolling 3D page with an editable model; it's popular-science narrative rather than a tool. The visuals are worth a look, but there's no reusable value for developers, making it a content piece. #3D#Web Frontend#Animation

    Yingxian Wooden Pagoda · A Millennium in Wood: a six-act scrolling 3D popular-science website with an editable model

    huangbai-AIPython★8

  • C minimalist-themes-for-hermes Eighteen light color schemes for Hermes Desktop, with an installable plugin, and the workmanship looks tidy; but it is ultimately just a set of themes, meaningless to anyone outside that desktop environment's users. #UI Design#Plugins

    Eighteen warm, light color themes for Hermes Desktop, with readable UI refinements and a simple installable plugin.

    mykeuraJavaScript★14

  • C brush-and-blade A browser fighting game with eight pixel characters, AI opponents and special moves, closer to a playable tech demo. There are many similar fighting demos, and nothing new appears in gameplay or engine, so the screenshots basically tell the whole story. #Games#Game Dev#AI#Animation

    An eight-artist pixel-art fighting game with AI opponents, special attacks, and animated sprite showcases.

    emollickJavaScript★9

  • C matrix-cli It adds a cascading Matrix digital rain effect to the Codex CLI and Claude Code terminals; this is ambient decoration, presumably simple to implement, with no practical help for daily coding—a good-looking but replaceable little toy. #CLI#Terminal#AI

    Layered Matrix-style code rain for Codex CLI and Claude Code

    onepunkJavaScript★7

  • C Paper-Reading-Skill A skill package for reading ML/CV/NLP papers, essentially a collection of prompts and workflows; anyone can write this kind of content and its effect is hard to verify, so a general assistant plus your own reading notes is better. #ML Research#Learning#Agent Skills

    Skill package for ML/CV/NLP paper reading

    sunny2109Language unknown★12

  • C tideglass-cove A small coastal scene built with Three.js and TSL, focused on atmosphere and rendering experiments. It looks nice as a technical-art sample, but there are no reusable components and no stated audience, so it isn't a tool. #3D#Web Frontend#Animation

    Tideglass Cove is a small coastal world designed to feel peaceful, vivid and alive. Built with Three.js+TSL

    iamtechartistJavaScript★11

  • C Codex-Proxy A standard proxy and process proxy manager for Windows, assembled from Tauri, Vue, C++, and .NET. Its name collides with OpenAI's Codex, its positioning is vague, and it never explains how it differs from existing process proxy tools. #Windows#Networking#Automation

    ProcWeaver: a Windows standard proxy and process proxy management tool, built with Tauri + Vue + C++ + .NET.

    jojhaaC++★8

  • C Crypto-Recovery Claims to scan over forty wallets, analyze encrypted vaults and recover seed phrases, wrapped in the language of forensics. Such tools are one step from wallet theft and lack any authorization boundaries, so it shouldn't be trusted or used. #Crypto#Secrets#Reverse Engineering

    Python-based cryptocurrency wallet forensics suite supporting 40+ wallet types on Windows, Linux, and macOS. Features encrypted vault analysis, wallet detection, seed recovery, multi-chain address der…

    juliannacorreiaPython★35

  • C ai-intro Workshop course materials from Posit for getting started with AI/LLMs; following along builds basic concepts, but it is essentially tutorial material rather than a reusable tool, and similar courses and official docs are already plentiful. #AI#LLMs#Learning

    A hands-on workshop introducing AI and LLMs

    posit-devJavaScript★6

  • C RecallOs-Ai Yet another long-term memory layer for large models, touting vector-plus-graph retrieval and local privatization, but giving only latency numbers and no architecture details; mem0 and Zep already hold this space, and no difference is visible here. #AI Agents#LLMs#RAG

    Give your AI agents continuous, persistent long-term memory with hybrid vector-graph retrieval — running 100% privately on consumer hardware with sub-12ms latency.

    raitulhPython★6

  • C MGS004NSM-MVS-U4-Bartop-Arcade A set of print files and structural reference documents for a homemade arcade cabinet, ready-made material for anyone wanting to build a wooden arcade cabinet; but it is a hardware drawing share rather than a software project, and there is no way to verify the design was tested on a real machine. #Games#Hardware#3D

    Print files & reference for my MVS-U4 inspired Bartop Arcade.

    multigamesystemLanguage unknown★13

  • C loom The title only says "simulating weaving," using Python for fabric structure simulation; the topic itself is interesting, but the repository has almost no explanation—no algorithm, no output examples, no usage—so it does not yet constitute anything usable. #3D#Science#Animation

    Simulated weaving

    nicolasloizeauPython★16

  • C hermes-grok-bridge It builds an SSH and native API bridge between Hermes and a Grok bot, but does not explain what Hermes is or how messages flow, and lacks a minimal example; chat bridge scripts like this are easy to write yourself. #AI Agents#APIs#Networking

    Hermes ↔ Grok Bot bridge: SSH relay and native API (:8642) methods

    shagghiesuperstarPython★10

  • C omarchy-blue-mosaic-theme A light theme for Omarchy, with a color palette and mosaic background texture. It's a personal aesthetic product, and comparable color schemes are easy to find, so it hardly counts as a new tool. #Linux#UI Design#Operating Systems

    A light Omarchy theme with a continuous mosaic floor and graphite-blue interface colors.

    ejuroCSS★10

  • C agent-stack-notes Four long Chinese articles plus three literature maps on building external memory systems for large models; the organization looks careful, but it is essentially a tutorial and notes collection that produces no code, so it is excluded under this list's criteria. #Learning#LLMs#Knowledge Base

    A tutorial and literature compilation for building external memory systems for large language models: four long-form Chinese articles (memory systems / seamless window switching / system checklist and…

    marikaguraLanguage unknown★20

  • C zuey-pi-setup It packages the author's own coding-agent setup (22 extensions) into a copyable snapshot with backup and restore scripts; useful to the author, but to anyone else it is just someone's dotfiles, with no general value beyond secret cleanup. #AI Agents#DevOps#CLI

    Portable snapshot of a pi coding-agent setup + backup/restore tooling. Clone to rebuild 22 extensions on another machine. Excludes secrets, cache, and machine state.

    mrgoonieShell★9

  • C CROWBRAIN A market-watching terminal for Solana and Robinhood Chain, selling points being a raven-style interface and rule-based signals; the signal rules are not disclosed, so real-trading value is doubtful, and on-chain monitoring dashboards are plentiful. #Crypto#Fintech#Blockchain

    Experimental market observation terminal for Solana and Robinhood Chain, featuring a crow-inspired interface, rule-based signals, and live engine logs.

    sopersonePython★15

  • C friday Claims to add a persistent memory layer to coding agents, solving cross-session amnesia. This space is already very crowded, and the repo is all slogans and emoji with no explanation of storage or integration details, so credibility is low. #AI Agents#LLMs#MCP

    ⚡ The open-source persistent cognitive memory layer for AI coding agents. Stop your Cursor, Claude & Copilot from having amnesia.

    itskieHTML★12

  • C Auto-Modeling-Skill A skill pack for the Huawei Cup math modeling contest, claiming to cover the whole competition workflow, but it lists no concrete capabilities and cites a launch date that hasn't arrived. Modeling scripts and templates like this are already everywhere. #Agent Skills#Learning#Science

    A Skill dedicated to the Huawei Cup mathematical modeling contest, aiming to be a one-stop toolkit covering the entire competition workflow; the first version launched on September 15, 2026

    weibu123Python★7

DSkip93
  • D GPUtw-Skill The official knowledge pack for GPUtw, teaching Claude Code, Codex CLI, Cursor and others to use its GPU cloud REST API. With 43 stars in 1.6 days, it looks more like vendor content marketing than an independent project; it has sample value as an agent-skill distribution paradigm but no independent technical or commercial value. #Agent Skills#GPU#SDKs

    GPUtw.ai official AI knowledge package — teach Claude Code, Codex CLI, Cursor, Copilot and Gemini CLI to use the GPUtw GPU cloud REST API

    GPUtw-aiTypeScript★43

  • D re4 Decompiles the GameCube version of Resident Evil 4 into byte-identical C/C++, laying groundwork for ports, mods and engine research. 51 stars in 1.5 days come entirely from decomp community reputation. There's no commercialization path, but it demonstrates strong reverse-engineering skill; the biggest risk is a copyright-holder DMCA takedown, with other decomp projects as peers. #Reverse Engineering#Game Porting#Games

    Resident Evil 4 (GameCube, G4BE08 debug build) — complete byte-identical decompilation to C/C++

    adonis-singhC++★51

  • D awesome-android-apps A curated list of open-source Android apps serving the need to find FOSS apps and de-Google devices. Its 66 stars in 1.4 days show the demand exists, but the list relies on community maintenance, heavily overlaps with F-Droid and existing awesome-android lists, and has neither technical barriers nor commercialization space. #Android#Knowledge Base

    a list of awesome open source android apps 📱

    fiedriPython★66

  • D HomeCar An Android Auto dashboard for Home Assistant displayed via a configurable full-screen WebView. Its 40 stars in 1.2 days come from the smart-home community; the scenario is clear but very narrow, Android Auto app review is strict with limited distribution, and it depends heavily on the Home Assistant ecosystem, making commercialization difficult. #Android#Self-Hosted#Automation#home-assistant

    Open-source Home Assistant dashboard interface for Android Auto with a configurable full-screen WebView.

    RenanEspinarJava★40

  • D openjev Attempts to reproduce Jev-like results on a single 3090, exploring the limits of running large-model inference on consumer hardware. Its 49 stars in 0.8 days come from hardware and local-model topics, but the description is vague and the goals and outputs are unclear, making it more an experiment log than a usable tool whose value is hard to judge. #ML Research#GPU

    Can we run something like Jev on a 3090 at home?

    TheoLeeCJPython★49

  • D vm2api Uses a virtual-machine emulation approach to turn subscription accounts into API calls, effectively extracting subscription quotas. Its 46 stars in 0.6 days reflect gray-market demand; the technique is automation plus anti-detection, clearly violates terms of service, can be banned anytime, cannot be commercialized, and carries extremely high legal and stability risks. #LLMs#APIs#Self-Hosted

    vm2api: subscription-to-API conversion using a fully simulated virtual machine approach

    dofasttedJavaScript★46

  • D text-humanizer Rewrites AI-generated text to evade detectors like Turnitin and GPTZero; 76 stars in 0.4 days shows de-AI-ifying is a real need. But it is fundamentally a cat-and-mouse game with detectors, accuracy can be countered anytime, and academic ethics and platform bans make it a gray area unfit as a standard dependency. #Content Writing#AI

    text-humanizer is an open-source project designed to convert text generated by AI to its "humanized" version. Bypasses the most of AI detectors such as Turnitin or GPTZero

    korcarcPython★76

  • D turkish-bank-logos-svg Collects original SVG logos of 35 Turkish banks and 5 payment institutions, solving the search problem for high-quality financial marks; 44 stars in 1.6 days from designers and niche local demand. It is a static asset collection with limited trademark licensing and little maintenance value, hard to commercialize. #Images#Fintech#UI Design

    Türkiye'deki 35 banka ve 5 ödeme / e-para kuruluşunun orijinal SVG logoları · Original SVG logos of 35 banks and 5 payment / e-money institutions in Turkey

    murattahtaciJavaScript★44

  • D awesome-rust-apps A curated list of Rust applications helping users discover desktop and CLI tools in the ecosystem. Its 49 stars in 1.7 days come from habitual Rust community attention, but awesome-rust and similar lists already exist, the list is highly homogeneous and maintained by enthusiasm, and there is no commercialization space. #Learning#Documents#Knowledge Base

    A curated list of applications built with Rust

    jaywcjloveRust★49

  • D picsio-mcp-plugin A vendor's MCP plugin for its own digital asset management, letting teams already on Pics.io connect assets to AI assistants. It's essentially a commercial integration with no reusable new method, and worthless outside that ecosystem. #MCP#Plugins#Images

    Pics.io MCP Plugin for AI Marketplaces (Anthropic, OpenAI, Copilot, Gemini)

    TopTechPhotoJavaScript★8

  • D Lazy_AMPR A Python script that compresses ampr game assets into LZ4, meaningful only in a specific handheld/emulator circle. The description is too brief, with no supported scope or risks, so outsiders can hardly judge whether it is usable. #Games#Performance#CLI

    compress ampr (or no ampr) games into LZ4

    NazkyPython★13

  • D aibuildai-llm-posttrain-agent Self-described as a recursively self-improving autonomous LLM post-training agent, it's all concepts with no reproducible interface or steps. A classic vision-first repo where it's impossible to tell what it can actually do today. #LLMs#AI Agents#ML Research

    AIBuildAI LLM-Post-Train Agent: a recursive self-improving (RSI) agent for autonomous LLM post-training

    aibuildai-incPython★27

  • D kutu-android-tv-guide A workflow write-up on using Claude Code with ADB to optimize old Android TVs, adding AirPlay and a lightweight desktop. It's a workflow note, not software you can install directly. #Android#Learning#Automation

    Claude Code + ADB workflow for optimizing older Android TV devices, adding AirPlay screen mirroring and building a lightweight custom launcher.

    bevcko16Language unknown★9

  • D obsidian-spaces It claims to add a "lens" layer to an Obsidian vault, but one sentence does not explain what the lens means or how to use it; the Obsidian plugin ecosystem already has many similar view and section plugins, and this lacks differentiation. #Knowledge Base#Productivity#Plugins

    A lens for your Obsidian vault

    psjdevTypeScript★9

  • D hyc A cybersecurity job-hunting roadmap plus resource list, essentially a content collection; many similar roadmaps exist, and phrasing like "strongest candidate" leans marketing. Beginners gain limited benefit, and it's hardly a new approach. #Learning#App Security#Network Security

    http://arcanum-sec.github.io/hyc/ - Hacking Your Career — a roadmap and verified resource library for becoming the strongest cybersecurity candidate you can be.

    Arcanum-SecHTML★11

  • D helloUI Self-described as a long-term design specification skill for responsive web and mobile products, with vague wording that doesn't reveal whether it's a usable skill, spec document or component library. Wait until it clarifies what it is. #UI Design#Agent Skills#Web Frontend

    Long-term design specification skill for responsive web and mobile products.

    cyMaiPython★11

  • D pase-omo Builds provider and workflow UI for Paseo, a rare agent platform, piling up terms like DAG graphs, real-time todos, and approval popups. The underlying platform itself is little known, so outside readers can hardly judge what it can actually run. #AI Agents#LLMs#UI Design

    OmO agent provider and workflow UI for Paseo - DAG graphs, live todo cards, approval popups, mobile-ready

    HakubisualTypeScript★7

  • D patch-devin Patches the Devin CLI with equal-length system prompts, essentially modifying a commercial tool's binary and prompts to bypass its validation. It's neither reusable technology nor easy to discuss publicly, so not recommended to follow. #Reverse Engineering#Prompting#LLMs

    Devin CLI equal-length system-prompt patcher (chisel / devin.exe)

    a0yarkPython★7

  • D book-answer Claims to distill classic works and mentor figures into an AI thought platform, but it stops at the vision stage without explaining data sources or concrete usage. Products for chatting with famous figures are already plentiful, and no new approach is visible. #AI#LLMs#Knowledge Base

    Book answer — an AI thought distillation platform for classic works and mentor figures (React 19 + Express + SQLite; user side supports Simplified Chinese / Traditional Chinese / English)

    helper15901590TypeScript★7

  • D RayCodes_Humanizer Another "make AI text sound human" skill pack claiming 100% humanization, when it actually just calls local Ollama to reword text, with unverifiable results. Similar prompts and rewriting tools are everywhere, the marketing tone is obvious, and 9 stars prove nothing. #Agent Skills#Prompting#LLMs#Content Writing

    Humanizer Skill 3.0: Make AI Writing Sound 100% Human Locally! - Local AI prose humanizer skill for Ollama, Claude Code, and Skills CLI

    47thtechcornerPython★9

  • D flex-writer-share A copywriting optimization agent listing a pile of "must-have" capabilities like web search, workflows, memory and multi-turn dialogue — which precisely shows a lack of focus. Similar agent templates are everywhere, and it's unclear which specific writing problem it solves. #AI Agents#LLMs#Prompting#Content Writing

    A copywriting optimization agent implementing essential agent capabilities such as web-search-tool, workflows, memory, multi-turn dialogue, and skills.

    CaptivePigPython★9

  • D numberwang A meme project using a small neural network to judge whether a number is Numberwang, purely community entertainment. Its 74 stars in 1.2 days come from HN-style humor spreading; it has no real use case or commercialization path and will die after the hype, with its only value as a neural-network toy tutorial. #AI#ML Research#Benchmarks

    A small neural network that decides whether a number is Numberwang.

    GraafHenkPython★74

  • D supvpn A repo of VPN/proxy reviews and free node lists, stuffing censorship-circumvention keywords to farm search traffic; 40 stars in 1.5 days mostly reflects SEO funneling, not technical recognition. No technical substance, clearly gray-market, may be purged by platforms anytime, and offers no investment or dependency value. #Networking#Privacy#Marketing

    Stable airport VPN reviews — personally tested and working: circumvention, scientific internet access, free nodes, free VPNs, free ss/v2ray/trojan nodes, Lantern, Google Play, circumvention VPNs.

    supvpnLanguage unknown★40

  • D cheatbench A repo under a security center's name, but the description is just a URL, with no explanation of what the benchmark tests, how to run it, or where the data is. Can't evaluate until docs are added, so not counted for now. #Benchmarks#LLMs#ML Research

    cheatbench.ai

    centerforaisafetyLanguage unknown★7

  • D wine-nx An undescribed Wine fork whose name suggests running Windows programs on the Switch, but it never explains how to build it or how complete it is, so there's no way to judge whether it's usable. #Game Porting#Linux#Windows

    danfromticoC★12

  • D awesome-enterprise-agent-guide Yet another awesome list, cataloging resources on building, governing and observing enterprise agents; list repos produce no tools themselves, their value rests entirely on future maintenance, and by this list's standards it isn't recommended. #AI Agents#Learning#LLMs

    Learn how to build, scale, govern and observe and enterprise grade AI agent using Gemini Enterprise Agent Platform.

    iam-veeramallaLanguage unknown★7

  • D navigate-business-situations A Codex skill pack that turns business strategy, workplace communication and career choices into prompts. It's a prompt collection whose effects can't be verified and with no code to inspect, so by this list's rules it isn't recommended. #Prompting#Agent Skills#Learning

    A Codex skill for navigating business strategy, workplace dynamics, communication, and career decisions.

    jianili-gtmLanguage unknown★10

  • D REPO-Ultimate-Assistant A trainer for R.E.P.O. with eighteen features like god mode and infinite resources; this is a game cheat, so under the list's rules it is scored low outright, without discussing how complete it is. #Game Cheats#Games

    Ultimate assistant for R.E.P.O. Features: God Mode, Infinite Health/Stamina/Energy, Infinite Scrap, No Weapon Cooldown, Customizable Stats, Money Hack, Freeze Enemies, Blind AI, Round Director + Credi…

    matsuirebeccaPython★7

  • D airpods_hotwheel_track The repo has no description, and the name doesn't reveal whether it uses AirPods head tracking to control a car or something else; the Swift code likely needs a real device to verify. Created a day ago with 9 stars and no usable information, readers can't judge whether it's worth opening. #Audio#Hardware#Games

    1204alSwift★9

  • D discord-crasher Framed around binary instrumentation and fuzzing findings, it actually points to a tool for crashing Discord. It offers no public research details and crosses the abuse line. #Exploits#Debugging#Reverse Engineering

    Some bugs found via binary instrumentation and fuzzing

    aftermathlabsRust★12

  • D arma-3-menu-mod Claims to be a shop system script for Arma 3 mission authors, but its naming closely resembles cheat repos from the same batch of accounts, so it's hard to rule out the same volume-farming. Not counted until the code is clear. #Games#Game Cheats#Plugins

    Script that enables ARMA 3 mission maker to easily implement a shop system in their custom mission.

    chri-cox970Language unknown★10

  • D Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED Packages a nonexistent "Qwen3.8-27B" as a BF16 main-weights plus NVFP4 hybrid deployment scheme, piling on jargon, with a QuickStart aimed only at a few machines like DGX Spark. Both name and description are untrustworthy, with almost no sign of real use. #LLMs#GPU#AI

    Uncensored Qwen3.8-27B Ultimate - BF16 master + NVFP4-MIXED deploy. QuickStarts for DGX Spark (one Dynamic DFlash lattice seat @ util 0.80), dual-Spark TP=2, RTX 5090 / PRO 6000. Serve with aeon-vllm-…

    AEON-7Python★9

  • D projeto-aviacao-anac A class project from an Alura data engineering course, running Databricks and SQL analysis on ANAC aviation data. It's coursework whose reproduction depends on the course environment, offering no incremental value to list readers. #Data Analysis#Learning#Databases

    Projeto da Imersão Engenharia de Dados — setembro/2026. Análise de dados da aviação brasileira com dados da ANAC, Databricks, Python e SQL.

    Alura-Imersao-Engenharia-de-dados-2026Python★11

  • D image2.5-api-prompt-gallery A dozen image-generation prompt recipes with usage examples, essentially funnel content for some API service. It's a prompt collection and doesn't count under the rules. #Prompting#Image Generation#APIs

    Twelve image2.5 (GPT-Image-2.5) prompt recipes with example outputs, model version, aspect ratio and copy-paste cURL, Python and JavaScript calls through the APIMart API.

    apimart-11Python★27

  • D game-dev-starter-kit An undescribed starter template for game development. Scaffolding projects are already homogeneous, and without documentation there's no way to tell what it does better than a template generator. #Game Dev#Learning#Build

    danicatGo★13

  • D aml-gnn-sar-generator No README, so one can only guess from the name that it uses GNNs to generate AML suspicious transaction reports; there's no usage, data or results, leaving outside readers unable to verify. 8 stars in under two days isn't even enough basis to watch. #ML Research#Fintech#AI

    Anubhavsingh311Language unknown★8

  • D qwen38-inference Claims to be a reasoning implementation and Triton kernels for Qwen3.8-27B, but that model doesn't exist and the repo offers no verifiable source; it's either a placeholder joke or stitched-together benchmark numbers, so skip it. #LLMs#ML Research#Benchmarks

    Qwen3.8-27B inference, CUDA/Triton kernels, reference baseline, and benchmark results.

    humanfiaPython★8

  • D pdf-translate No description; the name points to PDF translation. Such tools are already numerous, but the repository explains neither its technical approach nor how to use it, so there is no way to tell what it does better than existing tools. #Documents#LLMs#Productivity

    lxsssssssPython★14

  • D nvim-config The author's own Neovim config, self-described as bloated. It works for them but has little migration value for anyone else; if you want to copy a config, look at mature distributions like LazyVim instead. #Editors#CLI#Productivity

    my bloated nvim config

    vyrx-devLua★8

  • D dsh-opencode-go The description is a single line, 'make your DSH perfectly fit the opencode-go plan,' with no indication of what DSH is, what problem it solves, or how to install it. Such small Chinese repos are usually temporary config files that outsiders cannot use. #AI#LLMs#CLI

    Make your DSH perfectly compatible with the opencodeg-go plan

    DuskriverTypeScript★10

  • D belief-context-graph An undescribed research repo whose name suggests belief or context graphs, with content and usability unconfirmed. Wait for a paper or description. #ML Research#LLMs#Knowledge Base

    bigai-nlcoLanguage unknown★10

  • D developer-ai-writer AI copywriting for developers, with a rocket emoji in the title and claims of instant release notes and blogs. Vague, with no discernible difference from existing writing tools. #Content Writing#LLMs#Documents

    🚀 **An AI-powered writing engine for developers.** Automatically generate release notes, technical blog posts, documentation, and marketing copy in seconds. Streamline your shipping workflow!

    bookreaderpageLanguage unknown★7

  • D ringzero A Zig project with no description, and the README doesn't clarify its goal; it gathered thirty-plus stars within a day of creation yet explains nothing, looking more like an empty shell riding the language's hype, impossible to judge or recommend. #Operating Systems#Performance#Linux

    immanuwellZig★31

  • D leetcode_daily A personal problem-solving log, with Java solutions piled up day by day, without any tool or methodology distilled. It's practice notes with almost no reusable value for others. #Learning#Code Quality#Benchmarks

    Anubhavsingh311Java★8

  • D macos-nan-usage The description is only one line, "NaN Usage para macOS," so it's unclear whether it's a system monitoring tool or something else, and impossible to tell what it does or who it's for. #macOS#Monitoring#Performance

    NaN Usage para macOS

    conjderumbaSwift★6

  • D ServerManagementPanel A service management panel with no description. Self-hosted panels are already plentiful, and without stated features or differences there's no way to judge it. #Self-Hosted#DevOps#Monitoring

    coraspirinTypeScript★10

  • D harness-dio No description; the name points to some kind of harness or test framework, but there's no explanation of purpose or usage, so there isn't enough information to go on. #Testing#CI#Automation

    digitalinnovationonePython★7

  • D AI-AGENT-CERTIFICACION A Spanish-language collection of study materials on agent certification, with only materials and no code, producing no tool. By this list's criteria it's a pure content repository and is excluded outright. #Learning#AI Agents

    Material de estudio

    floridioracleLanguage unknown★6

  • D Awesome-Public-Key-Infrastructure-Platform A PKI platform list whose title is stuffed with emoji and "Star if you like it," heavy on marketing tone with entry quality unknown. It offers limited help to engineers actually building certificate systems and isn't worth the time. #Learning#Secrets#App Security

    Top Public Key Infrastructure (PKI) Platform (Opensource) 🌟 Star if you like it! 🌟

    ishandutta2007Language unknown★7

  • D eda-paper-writing There is no description; the name suggests something about EDA paper writing, but there is neither explanation nor tooling, and the twenty-odd stars gained in a day lack supporting content, so it is not recommended for now. #Science#Learning#Documents

    loujcLanguage unknown★24

  • D group-meeting-materials A lab group meeting materials repo, pure content with no code, of no use to developers outside the group. The description is honest but carries zero information; it's a personal or group data dump, not an open-source project worth viewing. #Knowledge Base#Science#Learning

    Group meeting materials for Tailin Wu's lab

    AI4Science-WestlakeULanguage unknown★8

  • D LunarEC One line, "Lunar CRM and ERP," with no feature description, screenshots or startup instructions, yet already flying the ERP big-platform banner. A typical empty placeholder repo that offers nothing usable today. #Web Frontend#Databases#Productivity

    Lunar CRM and ERP

    BekbolatBolebayTypeScript★7

  • D WinSux A PowerShell script repository with no README; from the name it looks like a collection of scripts to slim down or complain about Windows. It is impossible to tell which system settings it changes, so ordinary users should not run it rashly. #Windows#Automation#Productivity

    FR33THYFR33THYPowerShell★7

  • D .github Just an organization homepage repository introducing a rotating savings protocol on Stellar, with no runnable code or tools, so there is nothing for developers to evaluate. #Blockchain#Fintech#Crypto

    Organization profile for Susu Protocol — non-custodial rotating savings on Stellar

    SUSU-LABSShell★7

  • D zhihu-ai-cookbook Just a three-volume manuscript collection (AI Infra, RL, embodied AI), a content repo with no code or tools. Possibly useful for people who want to read the material, but it doesn't qualify as a candidate for this list. #LLMs#Learning#Content Writing

    Zhihu AI Blue Book · three-volume manuscript (AI Infra / RL / Embodied Intelligence)

    ZhihufrontierPython★11

  • D apis-that-find-your-next-clients A B2B lead API list that boasts of being one of the most valuable lists on GitHub, a resource list plus marketing that doesn't count under the rules. #APIs#Marketing#Jobs

    This GitHub repo is a powerhouse collection of lead-finding APIs you can start using immediately to fill your pipeline — from contact enrichment to B2B prospecting. One of the most valuable lead API l…

    cporter202Language unknown★7

  • D intro-repo A personal portfolio page written in Next.js by its author. It solves no problem and offers no reusable technical ideas, a typical practice project with no reading value for outside developers. #Web Frontend#Learning#Productivity

    It is my intro-repo. my portpolio. I made this project by Next.js

    eymenkaratas3347-dotTypeScript★10

  • D Soroban-Contract-Upgrade-State-Migration-Framework The name is long but there is no description, so it is unclear which specific contract-upgrade problem the migration framework solves; an anonymous account with zero explanation makes it impossible to assess whether it is usable. #Blockchain#Crypto#SDKs

    proverbs31ziggo-a11yRust★7

  • D The-Blood-of-Dawnwalker-Ultimate-Trainer A trainer collection for a pre-release game, full of emoji and entries like 'infinite health' and 'one-hit kill'; a typical cheat/trainer repository that scores low under the rules. #Game Cheats#Games#Windows

    🩸 Ultimate trainer & mod collection for The Blood of Dawnwalker. Infinite Health, Infinite Stamina, One-Hit Kill, Damage Multiplier, Set Game Speed, Edit Money, and more. Memory-only, no file modifica…

    Janne1157Language unknown★10

  • D yoursite A portfolio template named 'For Professional Portfolio Sites' with no description and no example site; a common placeholder repository. Similar templates are everywhere, and no difference is visible. #Web Frontend#UI Design#Productivity

    For Professional Portfolio Sites

    MP061codeLanguage unknown★8

  • D Marvels-Wolverine-PS5-Emulator-Offline-Trainer An offline Wolverine trainer riding the PS5 emulator hype, with tutorials and guarantees like 'zero ban risk.' This is game cheating and grey-market adjacent, so it scores low under the rules. #Game Cheats#Games#Windows

    🐺 Marvel's Wolverine on PC via KytyPS5 & SharpEmu emulators. Includes RakibOFC offline trainer with God Mode, Infinite Rage, One-Hit Kill. Full setup guide, system requirements, graphics tech breakdow…

    ModIndex-Game-Trainer-HubC++★10

  • D ai-automation-yield-pipeline A blueprint for building an automated revenue pipeline with LLMs and traffic distribution, full of marketing language and no runnable code, a typical content-funnel repo. #Marketing#Automation#LLMs

    A comprehensive technical blueprint and workflow guide for building automated, low-maintenance digital yield pipelines using LLMs, traffic routing, and AI automation. Stop trading time for dollars.

    bookreaderpageLanguage unknown★8

  • D Anthropic-Claude-Certified-Architect-Foundations-CCAR-F-Exam-Preparation-2026 A pile of exam prep material for some certification with keywords stuffed throughout the description, a typical content and SEO repo with no tool value. #Learning#LLMs#AI

    This repository provides study material and practice questions for the Anthropic Claude Certified Architect Foundations (CCAR-F) certification. It is intended for learners preparing for Anthropic exam…

    cooper-james-80Language unknown★9

  • D awesome-micro-saas-builder Another awesome list claiming you can build and ship a Micro-SaaS in minutes, but it's really a collection of links and templates. Repos like this go stale in three days and offer developers nothing actionable. #Marketing#E-commerce#Productivity

    A curated collection and automation engine for Indie Hackers to build, remix, and launch Micro-SaaS & digital products in minutes.

    travelmapworldLanguage unknown★8

  • D jevlike No description at all, yet it pulled 200+ stars in half a day, and the repo contents don't map to any clear purpose. The growth is anomalous; skip it. #AI#LLMs#Productivity

    vinnylarougePython★210

  • D Brave-Origin Titled "Brave Origin for free," it points to bypassing Brave's paid subscription, making it a crack/piracy item that is always scored low under the rules and not recommended for team or company environments. #Crack#Piracy#Privacy

    Brave Origin for free

    ChathurangaBWC++★7

  • D SANE The description is only the slogan 'Signal Above Needless Embellishment,' with no indication of what it provides or how to use it. An HTML repository with 6 stars, its purpose cannot be judged, so it can only be skipped. #Web Frontend#UI Design#Productivity

    Signal Above Needless Embellishment

    JTech-COHTML★6

  • D travel-guide-builder No README explanation at all, just an empty repository name. It is impossible to tell what it does, and there are no examples or usage, so it cannot be recommended to anyone. #Web Frontend#Productivity#Documents

    Lky20030520HTML★7

  • D WARDOGS-Ultimate-Trainer-Collection A game trainer collection mass-produced from the same account, with infinite health, added money, and free purchases, emphasizing single-player only. A typical cheat repository that scores low under the rules. #Game Cheats#Games

    🐕 Ultimate trainer & mod collection for WARDOGS. Unlimited Health, Add Money, Unlimited Ammo, Free Purchases, Instant Construction, and more. Memory-only, no file modifications. Single-player practice…

    ModIndex-Game-Trainer-HubLanguage unknown★18

  • D postiz-uploader The repository has only a name, no description and no statement of purpose; from the name one can only guess it relates to Postiz's upload step, so its content and usability can't be assessed and it isn't recommended for now. #Social#Automation#APIs

    gitroomhqPython★11

  • D vodiwalker-worker Also no description, just a worker name, so it's unclear whether it's a Cloudflare Worker or some other backend script. Not enough information to evaluate. #Productivity#Automation#Web Frontend

    VodiwalkerJavaScript★12

  • D avatar-speak-pro-download A repository named download, packaged as a TikTok advertising digital-human pipeline but actually more like a lead-generation download entry, with no implementation notes or license, so it is not recommended to open. #Piracy#Video#Marketing

    An automated AI avatar video generation pipeline designed to boost conversion rates for TikTok ads and Shopify funnels via text scripts.

    travelmapworldLanguage unknown★8

  • D dlss-5-amd-project The repo has no description at all, so its relationship to DLSS or AMD can only be guessed and usability can't be judged. The name rides a trend, so it doesn't count under the rules. #GPU#Hardware#Performance

    TheAutomaticC++★8

  • D aquarius The repo has no description, and the name doesn't reveal whether it's a tool or an app, so usability can't be judged. Wait for the author to add a README. #Productivity#Automation#Web Frontend

    VladHardAndSoftWareTypeScript★8

  • D mva No description and no judgeable content, just an abbreviated name, so purpose and usability can't be evaluated. #Productivity#Documents#Automation

    comarquetLanguage unknown★7

  • D kaggle_32gpu_free No description; the name hints at something related to free Kaggle GPUs, but it's impossible to confirm whether it's a script, a tutorial or traffic bait. With zero information it can't be assessed and is excluded outright. #GPU#Learning

    kumaresankpLanguage unknown★7

  • D C-Cpp-2026 An empty-description repository where only the name hints at a C/C++ learning code collection; with no information to judge by, it is a typical no-description, no-code, no-usage repo not worth opening. #Learning#Build

    luckymarkCMake★9

  • D dmghg-crk The repository has no description, and "crk" in the name hints at cracking; with no explanation and no reason to spend time verifying it, it is excluded outright under the rules. #Crack#Piracy

    newliver666Language unknown★8

  • D Passed-CCAO-F Exam-passing tips plus soft advertising for a training institution, a typical braindump lead-generation repository with no technical content, excluded outright. #Learning#Marketing

    Passed CCAO-F certifications on my first attempt. I prepared with regular revision, practice questions, and hands-on learning. SkillCertExams practice questions were also useful for testing my knowled…

    oliviajam-007Language unknown★7

  • D How-I-Passed-CCAR-P-Certification A typical certification advertorial: first-person 'passed on the first try' plus a link to an external question-bank site, with no code and no method. This is grey-market traffic bait and scores low under the rules. #Scams#Learning

    Just passed CCAR-P on my first attempt! I used SkillCertExams for some extra practice and it helped me feel more prepared.

    Emma-Parkar67Language unknown★8

  • D Passed-All-Anthropic-Certifications-on-My-First-Attempt First-person 'passed all Anthropic certifications on the first try' plus an external question-bank link, with no code and no method; a typical certification grey-market traffic bait that always scores low under the rules. #Scams#Learning#AI

    I’m happy to share that I passed all Anthropic certifications on my first attempt. Preparing for each exam helped me improve my understanding of Claude and practical AI concepts. I focused on practice…

    Juila-8809Language unknown★8

  • D multi-theft-auto-menu An MTA multiplayer cheat menu whose promo copy is full of high-performance and elite-level buzzwords. It's a game cheat, so directly excluded. #Game Cheats#Games#Scams

    Multi Theft Auto Inferno (2026) is the premier utility for advanced MTA multiplayer enhancement. Experience high-speed mod integration and fully optimized tools for elite server performance and gamepl…

    chase3478Language unknown★10

  • D crossfire-mod-attack A shooter game cheat repo with typical promotional copy and no technical explanation. Excluded. #Game Cheats#Games#Scams

    Looking to get the most out of your Crossfire experience? Attack mod is the perfect companion by your side!

    chloe9625Language unknown★10

  • D MA_playground A personal practice ground for learning mathematical analysis, with no description and no reusable tooling. It's personal notes, not something worth recommending as an open-source project. #Learning#Science#Documents

    I love mathematical analysis~

    xby474-devJavaScript★8

  • D Nala No description, and C++ plus a few dozen stars isn't enough to tell what it does. A high-star unnamed repo less than a day old; treat it as noise for now. #3D#Games#Performance

    yappologisticC++★34

  • D ip-studio The repository has no description and the README doesn't reveal what it does; there's no telling who it's for or how to start, so by the rules it's excluded outright. #Web Frontend#UI Design#Productivity

    gaoweiblue-sourceJavaScript★11

  • D jetbatang The repository has no description, and the name reveals neither its function nor its target users; with no usage or positioning, it is excluded under the rules. #Productivity#UI Design

    kyh0703Python★7

  • D windows-fps-boost-utility Yet another Windows "optimizer" claiming one-click frame boosts, with unverifiable precise figures like +35% FPS and −42% latency, but no implementation notes or code clues—a marketing page rather than a tool. #Windows#Games#Performance

    Boost your FPS, reduce latency, and optimize Windows for gaming — in one click. +35% average FPS, −42% latency, 98% system response.

    plumenodeblaze7Language unknown★11

  • D zolo The repository has no description and an unknown language, so there is no way to tell what it does or who it is for; a project less than two days old that cannot spare a single sentence about itself is not worth opening. #Productivity#Automation#CLI

    oriwellLanguage unknown★11

  • D pixel-and-ping No description, just TypeScript and single-digit stars, so readers cannot tell whether it is a pixel art tool or something else; empty-shell repositories like this should be treated as noise. #Web Frontend#Games#Social

    pixelpinggTypeScript★6

  • D Passed-CCAR-F-Exam Exam-passing tips plus promotion for a training institution, neither code nor verifiable material—a typical certification marketing repository. #Learning#Scams#Marketing

    Passed CCAR-F on my first attempt. Along with studying the core material, I used SkillCertExams to test my understanding with extra practice.

    robert-blake-960Language unknown★9

  • D system-one No description and no explanation; Python and a dozen stars are not enough to tell what it is for. An anonymous project less than a day old has nothing to evaluate. #Automation#Productivity#Self-Hosted

    snellingioPython★11

  • D domplus-apartment-project There is no description, and the repository name points to some apartment project, with no visible code value or reusable components, so readers have no reason to open it. #Web Frontend#Marketing#Social

    thekaroe-thailandLanguage unknown★8

Show HN 6

S

Show HN: Swift-Qwen3.8-27B, -58.3% thinking, x1.95 speed, accuracy of xhigh

Show HN

Post-training Qwen 27B with on-policy distillation that penalizes tokens tied to overthinking halves think length with almost no accuracy loss, directly cutting inference cost. It drew 80k downloads in 3 days plus community quantization and independent evals, strong momentum; monetization is API and licensing (only charging companies above $10M), but the risk is that base models' built-in effort control will absorb this value.

Hi everybody, we post-trained Qwen 3.8 27B to be more efficient by figuring out which tokens were linked to overthinking and penalizing them without "attacking" the reasoning length directly then fixe…

#LLMs#ML Research#Benchmarks#AI

▲26

A

Show HN: I built a router for agent tools

Show HN

OpenRouter for agent tools routes and calls tools from different vendors through one interface, fixing the fragmentation of rewriting code for every tool added. 67 stars plus Show HN hit a real standardization pain point; the path is take rates and marketplace distribution, with potential to become a standard gateway, but the risk is MCP official and ecosystems like LangChain building in the same capability.

#AI Agents#MCP#APIs#SDKs

▲10

A

Show HN: Give your AI agents access to WhatsApp

Show HN

Wraps WhatsApp as an MCP so agents can read, write, and search messages, with a Web UI for non-technical users, solving repeated WhatsApp integration per project; it drew HN discussion. Monetization is a hosted API/SaaS, but it directly violates WhatsApp's ToS, with high ban and data-compliance risk, and the official Business API is a strong competitor.

Hi, I built Chat-Man because I wanted a cheap way to give my agents access to WhatsApp without integrating a WhatsApp library separately in every project. You get WhatsApp MCP server, so you can conne…

#MCP#AI Agents#Social#Automation

▲14

B

Show HN: Fail2zig. A single-binary fail2ban replacement written in Zig

Show HN

A Zig rewrite of fail2ban replaces the regex engine with a bounded parser, persists state in SQLite, and writes nftables directly via netlink, addressing the risk of attacker-controlled logs triggering regex backtracking blowups. 0.4.0 just shipped with a website; HN discussion is cool but the direction is solid. Monetization is enterprise support or hosted protection; the risk is fail2ban's entrenched ecosystem and mature competitor CrowdSec.

Built fail2zig because I like fail2ban's log-to-ban model but wanted a smaller runtime with explicit limits on the work attacker controlled input can cause. fail2zig is a single static binary with bou…

#Network Security#DevOps#Self-Hosted#CLI

▲6

Hot New Repos

  1. saragordic/window-sweaters★424 A macOS menu-bar app that dresses your windows in knitted borders.
  2. FSECDEV/Threat-Intelligence-Hackers-Forums★329 A directory of hacker forums
  3. theoephraim/awesome-cloudflare-selfhosted★265 🍊☁️ Open-source* apps that replace a SaaS product, running in your own Cloudflare account
  4. Marcos66236/github-stars-history★227 Track and visualize the stars history of any GitHub repository. Open-source growth analytics and velocity tracking.
  5. vinnylarouge/jevlike★210
  6. anmolkapil/plexo★120 Speed up downloads by combining multiple network connections in parallel
  7. YongshengWin/VpsCT★91 A self-hosted server management panel for individuals and small teams, centrally managing VPS status, traffic, service configuration, and resource sharing.
  8. hirakujira/NEIN★84
  9. korcarc/text-humanizer★76 text-humanizer is an open-source project designed to convert text generated by AI to its "humanized" version. Bypasses the most of AI detectors such as Turnitin or GPTZero
  10. GraafHenk/numberwang★74 A small neural network that decides whether a number is Numberwang.

Star gain 1d

  1. debpalash/VoiceStudio+2821★33670
  2. affaan-m/ECC+1688★259083
  3. NationalSecurityAgency/ghidra+1218★77844
  4. Tencent/WeKnora+1210★24718
  5. eternity4719/HowToLiveBetter+1179★2132
  6. ai-sucks-butt/ai-sucks-butt+1170★1776
  7. addyosmani/agent-skills+902★94845
  8. heygen-com/hyperframes+700★50985
  9. Human-Agent-Society/reef+606★2581
  10. cursor/plugins+487★8192