tierlist.tech
中EN

10.22 Today's Picks

320 new projects found today, 194 rated.

Created 2026-10-20 – 2026-10-22Fetched 600Filtered 50Rated 38

S · Look today

S

php_uploadprogress_guard

HOSTED-POWERC

A small PHP extension that mirrors official fixes component by component across 5.6–8.5, steps aside once the engine ships its own patch, and lists deployment paths for Sury, cPanel, DirectAdmin and Docker — exactly what shared-hosting operators need this week. The cost is carrying an unofficial patch in production.

Protects an unpatched PHP against php2shell (CVE-2026-109618 / CVE-2026-109619): a small extension that mirrors the official fix component by component and steps aside once the engine carries it. PHP …

#App Security#Exploits#DevOps#Self-Hosted

★56

A · Worth watching

A

rl-env-use

browser-useHTML

Uses a Browser Use agent to mine real web pages for cases and auto-generate RL environments, solving the lack of real task distributions for agent training. Its 102 stars in 0.8 days ride the browser-use brand and RL training hype; the path is a paid API for environment generation as training data or evaluation; the risk is questionable environment quality and reproducibility, competing with in-house benchmarks.

Build RL environments for any domain from real cases that Browser Use agents mine from the web

#AI Agents#ML Research#Benchmarks

★102

A

kindi

ahtxPython

An inference engine written for old cards like the Quadro K420, running ternary-quantized 2B models at 20+ tokens/s with an OpenAI-compatible service. Squeezing usable speed from old hardware is a rare approach and the numbers are concrete; the limits are targeting just one card and one model.

A custom LLM inference engine for the NVIDIA Quadro K420: Ternary-Bonsai-2-2B at 20+ tokens/s with an OpenAI-compatible server

#LLMs#GPU#Performance

★26

A

opal

xingjian-baiPython

Has a coding agent continually rewrite the optimizer itself during training, close in spirit to automatic algorithm search like AlphaEvolve. Attractive to researchers working on training optimization, and 55 stars in a day shows some response, but there's only one sentence, so whether it installs and works depends on the code and experiments.

OPAL: a reactive optimizer. A coding agent rewrites the optimizer while your model trains.

#AI Agents#ML Research#LLMs#Automation

★55

A

ps5-bazzite-deck

VivaLaVentShell

Brings Bazzite's Steam Game Mode image to the PS5, with Game Mode, VCN hardware decoding, and Wi-Fi all working. It is real progress for people tinkering with PS5 Linux; the limits are coverage of only specific firmware and models, with a high barrier for ordinary players.

Bazzite Deck (Steam Gaming Mode) images for the PS5 with working Game Mode, VCN and Wi-Fi

#Linux#Games#Hardware#ps5

★19

A

video-to-blender

gokayfemPython

Converts reference videos or images into editable Blender procedural node trees, with a tested library of Geometry, Shader and compositing nodes, packaged as a coding-agent skill. People recreating 3D references save a lot of node-building time; the limits are dependence on a specific agent client and 16 stars with no verified results.

Turn a reference video, image or idea into an editable procedural Blender node system: a coding-agent skill with a tested Geometry/Shader/compositor node-group library

#3D#AI Agents#Agent Skills#Animation#blender

★16

A

HistorianRCA

ScottDuncanAILanguage unknown

A root-cause-analysis benchmark built from 90 days of simulated chemical-plant historian data, with faults buried for users to locate. Industrial time-series RCA has long lacked a public reproducible dataset, so process-control and time-series diagnostic teams should take a look; limits are a single scenario and small scale.

HistorianRCA: a root cause analysis benchmark. 90 days of simulated plant historian data from a steam-heated hold tank, with hidden faults to find and diagnose.

#Benchmarks#Data Analysis#Science

★17

A

matinee

Jarvis322Kotlin

A quiet, ad-free launcher for Google TV and Android TV, small in scope with an honest description, and TV launchers are indeed a long-standing pain point. Mature alternatives like Projectivy exist, so it will stand or fall on polish.

A calm, ad-free home screen for Google TV and Android TV

#Android#UI Design#Productivity

★42

A

freecad_solidworks_plugin

djbarrowPython

Lets FreeCAD open SolidWorks SLDPRT/SLDASM and Parasolid files directly, saving a paid conversion step. Mechanical design has long lacked this path and the scope is clearly stated; the open question is fidelity on complex assemblies, which nobody has verified.

FreeCAD addon: open and import SolidWorks parts (.SLDPRT), assemblies (.SLDASM) and Parasolid files

#3D#Plugins#solidworks

★14

A

php2shell-lab

mikensecShell

Packages offline reproduction of a PHP RCE vulnerability and the detection rules written from it into a rerunnable Shell lab, explicitly excluding exploits and requiring your own samples, which is directly useful for blue-team detection engineering. The scope is restrained and the description honest; the limit is that it covers a single CVE and you still have to find samples yourself.

Offline, reproducible lab to safely analyse an untrusted PoC for CVE-2026-109618/109619 and build detection rules. No exploit included; bring your own sample.

#Exploits#App Security#Network Security#Testing

★19

A

OpenSWOS

benbaker76C

Rewrites Amiga's Sensible World of Soccer in C as a cross-platform, network-playable open-source version; classic game remakes are always a community need. The risks are copyright on original assets and doubts about completeness for a repository one day old.

An open-source C port of Sensible World of Soccer (Amiga), for Windows, macOS, Linux, Amiga and the web, with online multiplayer.

#Games#Game Porting#WebAssembly

★16

A

toolkit

SliverkissGo

A pure Go middleware that adds tool-calling capability to any LLM API, so Go backends can use function calling without binding to the Python ecosystem. The positioning is clear in one sentence and the scope restrained; the challenge is the ecosystem advantage of mature gateways like LiteLLM.

A pure Go middleware library that adds tool-calling capability to any LLM API.

#LLMs#APIs#SDKs

★12

A

quiet-tizen

jss8649Python

A native, clean launcher and reversible debloat tool for the Samsung U8000F/Tizen 9, with notes written for agents to read. Targeting one specific model feels honest, but it only works on that TV and touching system partitions is risky.

Native Quiet Home launcher, reversible cleanup tools, and an agent-readable guide for the Samsung U8000F running Tizen 9.

#Embedded#Hardware#Privacy#tizen

★64

A

torrentplayer

torrentplayerRust

A desktop torrent player written in Rust that streams while downloading and seeks in seconds, explicitly offline with no account or telemetry. It's genuinely convenient for people who watch torrented video often, but Stremio and WebTorrent already cover similar needs, and stable decoding is the real test.

Stream any torrent, video or audio, and skip to any moment in about a second. Free, open-source desktop app for Windows, macOS and Linux. No servers, no accounts, no telemetry.

#Streaming#Video#Downloaders#torrent

★27

A

ComfyUI-Ming-3-tts

Saganaki22Python

Wraps Ming-3.0-tts as a ComfyUI node so video dubbing users can generate speech directly in their workflow. The scope is restrained and the description honest, but it is essentially a port whose ceiling depends on the upstream model, with similar TTS nodes as alternatives.

Native ComfyUI port of Ming-3.0-tts, supporting 0.5B and Tiny 9B speech generation.

#Speech#Plugins#comfyui

★17

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

BHas a spark50
  • B dyt-ashare-cli Makes A-share market research a reproducible CLI with agent skills, solving non-reproducible quant scripts and inconsistent data conventions. Its 301 stars in 1.6 days ride the Claude Skill ecosystem traffic dividend; monetization could be data subscriptions or institutional research terminals; the biggest risks are data-source compliance and akshare/tushare already owning developer mindshare. #Fintech#CLI#Data Analysis#Agent Skills

    DYT CN: reproducible China-market research CLI and agent skills

    renjiegodRust★301

  • B multicam-recorder Controls multiple Android phones from a desktop browser to record 4K in sync, with face-priority focus and upload-while-recording, genuinely useful for multi-camera talking-head and livestream work. The feature description is specific but long, and phone models and Wi-Fi prerequisites are under-explained. #Video#Android#Web Frontend

    Multi-cam recording: a free open-source multi-camera phone recording tool that controls multiple Android phones to record in sync from a computer with one click. Live view, individual zoom, native pho…

    dean9703111TypeScript★15

  • B rozklad-mcp It turns real-time schedules, delays and fares for Polish trains, coaches and city transit into an MCP service, actually connecting to data rather than wrapping another model layer, directly usable by people building European travel agents; the risk is scraping stability, and the emoji in the description are a bit noisy. #MCP#AI Agents#APIs

    🚆 Rozkład jazdy PKP, PKS i komunikacji miejskiej dla AI – serwer MCP: wyszukiwarka połączeń z przesiadkami, odjazdy, opóźnienia na żywo, ceny biletów. Polish train & bus timetables for Claude, ChatGPT…

    ContentWritercoTypeScript★14

  • B motion2code Has coding agents rebuild motion videos frame by frame into SVG plus GSAP code, closing the translation loss from design mockup to runnable animation. It stands out among similar skills through a verifiable frame-by-frame check loop; the path is design-tool plugins or motion outsourcing efficiency, but Lottie/Rive already cover most cases and the skill is easy to copy. #Agent Skills#Animation#Web Frontend#AI Agents

    Skills for coding agents: rebuild a motion design video as looping SVG + GSAP code, checked frame by frame against the source

    lovisdotioHTML★155

  • B PACE-Bench A benchmark evaluating coding agents building from PRDs in mature enterprise repositories, hitting the debate that AI writes code faster than it can be reviewed. The methodology is clearly explained, but paper companion code usually waits for data release, so for now only the ideas are visible. #Benchmarks#AI Agents#Code Quality

    Official code and project page for PACE-Bench: When Coding Agents Outpace Review: Evaluating PRD-Driven Development in Mature Enterprise Repositories.

    TianqiLiu1Python★17

  • B rigup Turns a gaming PC or Mac into a local AI server in one click, with a web panel to install and switch models, explicitly targeting 12–24GB Nvidia cards and Apple Silicon. The positioning and hardware scope are honestly stated, but it goes head-on against Ollama and LM Studio, and 18 stars show no differentiation yet. #LLMs#Self-Hosted#Image Generation#macOS#GPU

    Turn a gaming PC or a MacBook/Mac Mini into a local AI server with one install. A web dashboard installs and switches the models: chat and coding on both, image and video on NVIDIA. Tuned for 12 to 24…

    guscanalesPython★18

  • B ios-icon-mcp An MCP service that organizes iPhone home screen icon layouts through any MCP client. It is small but clearly explained, and people automating desktop organization would want to try it; the limitation is reliance on bridges like iOS Shortcuts, and 20 stars mean no real feedback yet. #MCP#iOS#Productivity#Automation

    Organize your iPhone home screen from any MCP-compatible AI client

    huangzheng2016Python★20

  • B home-tv-cn An Android TV launcher customized from Aerial, focused on entry points to popular mainland Chinese video apps — a real need, described with restraint. But it is essentially a themed fork, and keeping in sync with upstream long-term is the big question. #Android#UI Design#Video

    Jianju TV: a minimalist Android TV launcher for commonly used video apps in mainland China, customized based on Aerial.

    Jimu888Kotlin★49

  • B pi-stretchtext It brings stretchtext into a chat interface: the model gives a one-to-three-sentence answer, and phrases you care about expand in place into longer text that can expand further — a rare bit of interaction novelty, not another wrapper. The limits are that it requires Pi, and whether information density actually improves is unproven. #LLMs#UI Design#Web Frontend

    Stretchtext replies for Pi: the model answers in one to three short sentences, and the phrases that hold more open in place into longer text, which may hold further phrases to open.

    MonadicMikeTypeScript★12

  • B space-handdrawn-motion-video Turns whiteboard hand-drawn-style animated videos into an Agent skill: handwriting, marker line art, and frame-by-frame MP4 rendering synced to voiceover. It saves editing work for explainer-video creators, but skill repositories are heavily homogenized and actual output quality is unverified. #Video#Animation#Agent Skills

    Whiteboard hand-drawn style motion video Skill: handwriting, marker line art, taped screenshots and video clips, voiceover timing, beat-synced music, rendered frame by frame into MP4

    SpaceZephyrJavaScript★21

  • B globes A real-time glass water ball simulated with WebGPU and three.js, a pure visual demo. Anyone learning WebGPU refraction and fluid shaders can read the code directly; the scope is restrained and honest, but it only shows an effect with no reusable library or docs. #3D#GPU#Web Frontend#Animation

    Glass globes of water, simulated live on the GPU with WebGPU and three.js

    willwaltzTypeScript★27

  • B StreamingSwap The official implementation of real-time video face swapping, with streaming processing a step beyond offline swapping, so research value is clear. But the repository only offers code and a paper link, without inference cost or usage barriers, and the face-swap space is already crowded. #Video#ML Research#AI

    The official code of StreamingSwap, a real-time video face swapping method.

    aim-uofaLanguage unknown★21

  • B ComfyUI-MOSSImageTextEncoder A ComfyUI node that passes reference images through a vision tower and injects them into text prompts, aimed at people needing MOSS-Image conditioning. The implementation path is described concretely, but it depends heavily on specific weights, so ordinary users cannot easily tell whether it will run. #Image Generation#Plugins#comfyui

    ComfyUI node that feeds a reference image through the Qwen3.5-9B vision tower and injects its embeddings into the text prompt (MOSS-Image conditioning)

    ThetaCursedPython★19

  • B turnover A macOS screen reader running alongside VoiceOver; accessibility is indeed an under-served direction and worth acknowledging. But the repository is only a day old, with no demo or compatibility notes, so completeness cannot be judged—it looks like an early attempt. #Accessibility#macOS#Audio

    A screen reader for the Mac that works alongside VoiceOver. Free and open source.

    TurnOver-For-MacSwift★12

  • B EvaOmni An evidence-grounded benchmark for audio-visual reasoning that requires models to give verifiable support rather than just answers. The direction is reasonable, but multimodal benchmarks have appeared densely over the past two years, and without data scale or leaderboard results the increment is hard to judge. #Benchmarks#Multimodal#ML Research

    An Evidence-Grounded Benchmark for Audio-Visual Reasoning

    VisionXLabPython★19

  • B php2shell-bypass-aslr-exp It claims to exploit without relying on display_errors and memory_limit and without modifying the official version, which if true advances exploitation conditions; but 419 stars in half a day looks more like CVE hype than verified quality, and the README doesn't explain the ASLR bypass approach. #Exploits#App Security#Linux

    A php2shell exploit that requires no modifications to the official version, and does not require display_errors or limit_memory.

    9biePython★419

  • B php2shell-scanner A non-destructive scanner that uses crt.sh and subfinder for subdomain discovery, then grabs banners to infer versions, suited to quick asset inventory. But version detection has a high false-positive rate, and this is the fourth scanner for the same CVE this week. #App Security#Exploits#Network Security#CLI

    Non-destructive scanner for CVE-2026-109618 / CVE-2026-109619 (PHP core RCE). Subdomain recon via crt.sh + subfinder, robust banner grabbing, version-based detection.

    JoasASantosPython★50

  • B php-cve-2026-10961x-lab-poc Claims to reproduce an unauthenticated RCE in PHP 8.5.11 and contrasts it with the fixed version, written like a proper vulnerability experiment. But multiple repos with the same ID appeared within two days, and the ID-version pairing is hard to verify, so treat it as an untrusted PoC for now. #Exploits#App Security#php

    PHP CVE-2026-109618/109619 — self-contained unauthenticated RCE on stock PHP 8.5.11 (in-band PS(id)→cookie memory disclosure + UAF chain), full control ladders, fixed-version negative control on 8.5.1…

    dinosnPython★49

  • B mcsmart-butler A local panel built on McDonald's official MCP for single-person calorie-budget meal planning, group orders and transparent bill splitting, with a deterministic core doing the math, which is the right call. But it is a contest entry locked to one chain and one region, so extensibility is limited. #MCP#AI Agents#E-commerce

    A local panel and Agent Skill based on McDonald's official MCP: single-person calorie/budget meal planning, multi-person group ordering with transparent bill splitting, group-buy price comparison, and…

    HanmoyTypeScript★23

  • B reader-simulator It lets you try each firmware version of the Xteink X3/X4 in a browser or on Windows before deciding whether to buy, a smart tool for a narrow need, and it honestly labels itself unofficial. But the audience is limited to people considering this one e-reader. #WebAssembly#Hardware#Windows

    Try Xteink X3 / X4 Classic / X4 Pro e-reader firmwares without the reader: in the browser or on Windows (unofficial)

    PewPewzxcPython★13

  • B TF2-HUD-Studio A cross-platform TF2 HUD development tool that turns HUD editing from hand-written files into visual editing. The audience is clear but narrow, the HUD community already has mature editors, and the project does not explain its key differences. #Games#Editors#tf2

    A CrossPlatform Studio for TF2 HUD development, Perfect for everyone

    XeraFang-SolutionsRust★13

  • B local-autoresearch Runs an automated research loop locally on a MacBook for geometry, scheduling and ML experiments, and the author's direction is credible. But the repo is only half a day old with a one-line description, so what Flint is and how much can be reproduced remains unclear. #ML Research#AI Agents#Science

    Local autoresearch on a MacBook, with Flint: geometry, scheduling and machine learning experiments.

    conglu1997Python★18

  • B open-sesame A web penetration testing skill for Claude Code that follows the OWASP WSTG process and outputs PDF reports for developers and decision-makers. The process is described concretely, but similar AI pentest skills are already dense, it does not say how to verify result credibility, and 15 stars lack real usage evidence. #AI Agents#Agent Skills#App Security#Testing

    🔓 An AI web-pentest skill for Claude Code. Point it at a site you're authorized to test — it maps the attack surface, works through the OWASP WSTG, validates findings, and writes dev + stakeholder rep…

    hassan-algoHTML★15

  • B CinderVM Claims to be a deterministic bytecode VM for agent execution, with image verification, snapshots and hash-chain replay logs. The problem is fresh, but 65 stars seven hours after the repo was created and a one-line description show nothing runnable, so suspect agent hype-chasing first. #AI Agents#Operating Systems#Testing

    Deterministic bytecode VM for agent execution - verified images, snapshots and a hash-chained replay journal

    isabekastsucRust★65

  • B agent-microvm An AWS microVM isolation sandbox for agents, with a Python SDK, CLI and MCP server. The need is real, but E2B, Daytona and others already occupy the space, this looks like a thin wrapper, and 44 stars in half a day lack usage validation. #AI Agents#MCP#SDKs#DevOps

    Python SDK, CLI, and MCP server for isolated AWS microVM sandboxes

    jnh12Python★44

  • B hermes-cua-spaces Runs Hermes Agent's terminal, file, and code tools inside Cua Space. It's an integration within the trycua ecosystem with a sensible direction, but it only matters to people already using Cua, and 24 stars show no adoption yet. #AI Agents#Automation#MCP

    Run Hermes Agent's terminal, file and code tools inside a Cua Space

    trycuaPython★24

  • B Champagne Runs Mac OS X 10.6-era apps on other platforms, implemented in C++. The approach resembles Darling but targets older binaries, making it worth a look; the limitation is a one-line description, so whether it actually runs is unknown. #macOS#Reverse Engineering#Operating Systems

    Run applications from Mac OS X 10.6 Snow Leopard on anywhere.

    unixziiC++★38

  • B XGB430vProDigiDecrypt Targets Spain's Digi XGB430vPro router: logs in, pulls the config, and decrypts admin and PPPoE/SIP passwords. The scope is very narrow but clearly explained, usable as-is by users whose ISP locked them out of admin; the limitation is support for only one device, so other models require your own modifications. #Reverse Engineering#Secrets#Privacy#Hardware

    Script Python que inicia sesión en tu router XGB430vPro de Digi, se descarga la configuración, la descifra y te muestra las credenciales de administrador, PPPoE y SIP.

    vvalencianPython★13

  • B php2shell-scan A read-only, pure-standard-library php2shell reachability probe and change monitor, the most measured of this batch of scanners; but version-based judgments inevitably produce false positives, and four or five similar tools appeared this week, so it's not worth separate time. #App Security#Exploits#Network Security#CLI

    Reachability + change detector for the php2shell PHP RCEs (CVE-2026-109618 / CVE-2026-109619). Agent-ready, stdlib only, read-only by default.

    0xrsydnPython★13

  • B WebSocket4D A WebSocket client and server library for Delphi and FreePascal, an ecosystem that genuinely lacks ready-made options, and the description is plain and accurate. But it is only a protocol binding, with no visible test coverage or documentation depth, so stability must be judged from the source. #Networking#Windows#SDKs

    A Windows WebSocket client and server library for Delphi and FreePascal.

    DelphiMan68Pascal★14

  • B vlc-redesign An unofficial VLC 4 Windows build using Juxtopposed's CC BY interface, honestly stating it is not affiliated with VideoLAN. But it is only a binary distribution, with no notes on changes from upstream and no way to verify the build's provenance. #Video#Windows#UI Design

    Unofficial VLC 4 build with the interface designed by Juxtopposed (CC BY 4.0), running on Windows. Not affiliated with VideoLAN.

    KuchiSoftsC★13

  • B ShaolinMonks-Android It uses XboxXBE-Recomp to statically recompile Shaolin Monks to Android, with Vulkan, touch controls and graphics options, and explicitly ships no game files; technically interesting, but tied to a single 2005 game, with legal risk always hanging over it. #Game Porting#Games#Android#Reverse Engineering

    Unofficial Android (arm64) port of Mortal Kombat: Shaolin Monks (Xbox), statically recompiled with XboxXBE-Recomp. Vulkan renderer, touch controls, gamepad support, optional CRT/scanline filters, FXAA…

    RanggaCasperC++★19

  • B Kumquat A macOS offline file converter where you drag files and drop them onto a radial wheel via shortcut to convert, a bit of interaction cleverness. But the converter category is crowded, and the description covers only the operation, not which formats are supported, so the trade-offs are invisible. #macOS#Productivity#Images

    Офлайн-конвертер файлов для macOS: перетащите файл, нажмите клавишу, бросьте на колесо

    aeroboss13Swift★13

  • B Granny-Benchmark Probes model values with a trolley-problem-style question (would AI let an elderly person die over copyright), clickbait-titled but pointing at a real issue. Unfortunately it is a one-off mini-evaluation with no sample or scoring methodology, more entertainment than research value. #Benchmarks#LLMs#ML Research

    Would an AI let a grandma die to protect copyright?

    ThioJoeJavaScript★14

  • B UnblockMCDownload One-click patching of PCL/HMCL exe files to restore downloads, addressing the real pain of domestic launchers failing to fetch resources. The approach is binary patching, so it breaks whenever upstream updates, and several similar projects appeared around the same time. #Games#Windows#Downloaders

    An app that one-click modifies the exe of PCL, PCLCE, HMCL to restore download functionality

    TianMengLuckyRust★12

  • B eze Upload an HTML folder and get a private site accessible on the team intranet, with a database and uploads. The idea is practical, but one sentence packs in a database, real-time updates, AI, and "open-source Shopify Quick," blurring the boundaries and making completeness hard to judge. #Self-Hosted#Collaboration#Web Frontend

    Self-hosted private websites for your team. Upload a folder of HTML and get a link only your coworkers can open, with a database, live updates, uploads and AI built in. An open-source take on Shopify …

    ahyieldGo★14

  • B be-ambitious-harness Adds two acceptance gates and an independent verifier to Claude Code to stop agents from grading their own work. The idea is simple and effective, but it is only a pure-markdown prompt convention whose effect depends entirely on model compliance, and similar skills are everywhere. #AI Agents#Agent Skills#Prompting

    Make AI agents aim higher without losing control: a Claude Code skill with two bars per task, a cheap gate check before expensive work, and an independent verifier that never sees the worker's reasoni…

    alexpiarPython★22

  • B grids A hand-written Windows multi-pane file manager in Rust and Win32, aiming for fast startup and operation. The positioning is clear, but the space is crowded from Total Commander to Files, and no key differentiator is stated. #Windows#Productivity#Performance

    Grids — fast multi-pane file manager for Windows (Rust, Win32)

    efwwRust★16

  • B CVE-2026-109618-php2shell-lab It packages the PHP session.upload_progress UAF into a Docker reproduction environment and PoC, somewhat useful for defenders verifying the vulnerability; but a dozen same-topic repos appeared this week, and it offers nothing beyond a container, with no patch comparison or exploitation conditions. #Exploits#App Security#Network Security

    PHP2Shell (CVE-2026-109618 / CVE-2026-109619) Docker lab and PoC. Use-After-Free in PHP's session.upload_progress.

    0x4148Python★15

  • B recast-video-dgx It breaks video character replacement into shot references, MiniMax H3, dual DGX Spark compute and dynamic quality checks, a more concrete pipeline than most 'AI video' repos. But the hardware bar shuts out most people, and the tutorial can only be followed verbatim. #Video Editing#AI#Multimodal

    Video character replacement production Skill: storyboard and per-shot references, MiniMax H3 ComfyUI, dual DGX Spark joint compute and multi-machine segmentation, original audio synthesis and dynamic …

    JayceHuangPython★18

  • B SunoCleanser It cleans up quality artifacts and noise in SUNO-generated audio, a concrete pain point for AI music creators. But the description is a single 'tries to', with no before/after samples and no word on whether it denoises or regenerates. #Audio#AI#ML Research

    An app that tries to minimize SUNO's artifacts and noises

    MeFuMoPython★12

  • B ps5-linux-1240-build A Linux boot package for PS5 firmware 12.40, heavily overlapping with several PS5 Linux projects from the same period. The description does not clarify whether these are its own changes or a repackaging of others' work, leaving provenance and maintainability in doubt. #Linux#Hardware#ps5

    Working Linux Loader For PS5 Firmware 12.40 Phat

    Wigglez-sudoLanguage unknown★12

  • B Quant-research-lab-v0.1.0 Industry ETF rotation plus multi-factor backtesting, with trading-cost accounting and bilingual Chinese-English reports, the workflow is complete; but quantitative backtesting frameworks are crowded, and the version number stuffed into the repo name smells like a personal research draft. #Fintech#Data Analysis#Benchmarks

    Python framework for monthly sector ETF allocation and multifactor stock backtesting, with transaction-cost accounting and separate Chinese and English research reports.

    ChungKit-WANPython★35

  • B PCL-Helper A single-file WPF tool restoring PCL download functionality, overlapping with another exe-patching project from the same period. It is genuinely useful to domestic Minecraft players, but relies on modifying a third-party client, so maintenance costs are high and its lifespan short. #Games#Windows#Downloaders

    A tool to restore PCL download and modpack installation functionality, native WPF single-file GUI.

    Turing-MirrorC#★16

  • B json-const-content-mapper A TypeScript content mapper that infers const types from JSON imports. It solves a very narrow type-inference problem and is cleanly written, but it does not state its use cases, and for most people satisfies or as const is enough. #Web Frontend#Build#Testing

    A TypeScript content mapper that types JSON imports as const

    jaydenfyiTypeScript★12

  • B Snake-Reyers An interactive Snake game for TikTok livestreams where viewers control it via chat. It is a usable livestream interaction toy but serves a very narrow audience, depends on unofficial platform interfaces that may break at any time, and similar projects are already common on GitHub. #Games#Streaming#Social#tiktok

    Sistema gratuito para lives interativas de cobrinha no TikTok

    golgol2Python★13

  • B content-team-agents Splits PM, engineering, QA, and growth into Claude Code sub-agents, plus YouTube/Instagram skills to run content channels. It is a prompt-orchestration template whose effectiveness depends heavily on personal workflow, and similar repositories are already everywhere. #AI Agents#Agent Skills#Content Writing

    Claude Code agent team (PM, Scout, Engineer, QA, Growth) plus YouTube and Instagram skills for running a content channel with human approval.

    TifisCodeLanguage unknown★23

  • B kiosk-satellite-dashboard A Home Assistant panel for Kiosk Satellite on Echo Show/Spot, pure HTML/CSS/JS. It's cleanly built, but it's essentially a personal dashboard config, more useful to copy and adapt than to use directly. #Self-Hosted#Web Frontend#home-assistant

    Home Assistant dashboard for Echo Show and Echo Spot running Kiosk Satellite. Plain HTML, CSS and JavaScript.

    sylvertomJavaScript★16

  • B szt A macOS command-line sizetree tool for seeing what's eating disk space, distributed via Homebrew. It's probably cleanly written, but ncdu, dust, and gdu already cover this thoroughly, so there's no reason to switch without a distinctive interaction model. #CLI#macOS#Performance

    Source code of Mac CLI tool szt (sizetree), which shows what's taking up your disk. Distribution via Homebrew

    urttiSwift★16

  • B php2shell A PoC plus vulnerable lab claiming a PHP session.upload_progress pre-auth RCE. If the CVE is real it's worthwhile vulnerability research, but the ID and year look self-invented and 37 stars in half a day looks like hype-chasing; better to wait for an official CVE record before spending time. #Exploits#App Security#Network Security

    CVE-2026-109618 / CVE-2026-109619 — PHP session.upload_progress RCE PoC & Vulnerable Lab

    wakemeiflosePython★37

COrdinary66
  • C day-trader-skill Splits the day-trading workflow into 11 Claude skills, solving retail traders' manual work of watching charts, chasing news, and reviewing. Its 234 stars in 1.4 days ride Claude Skill hype and get-rich narratives rather than a validated strategy; commercialization is limited to subscriptions or signals with a low ceiling and investment-advice compliance issues; risks are no real alpha and platforms banning financial skills. #AI#Agent Skills#Fintech#Prompting

    11 Claude skills for day trading: scan thousands of charts, track breaking news, time every entry and exit, and replay every outcome before you risk a dollar.

    JakeschincariolPython★234

  • C php2shell-poc A PoC reproduction of a new PHP session.upload_progress CVE, helping security teams quickly verify exploitability; 921 stars in 0.7 days is purely disclosure-window squatting. Monetization can only come from folding it into a scanner or threat intelligence, not from the repo alone; the risk is that interest fades within two weeks and it may be taken down for abuse. #Exploits#App Security#Network Security#Monitoring

    CVE-2026-109618/109619 — 5 RCE chains + detection rules (Suricata, Sigma, ModSecurity). PHP session.upload_progress pre-auth exploit research.

    ikowPython★921

  • C open-bots An out-of-the-box open-source agent collection with search, sandboxing, file analysis, and payment registration, solving the tedium of building an agent toolchain yourself. But similar all-in-one bundles have flooded in recently, so 89 stars in 1.1 days is unremarkable; monetization could be a hosted version or enterprise private deployment; the risk is head-on collision with Omnara, OpenHands, and others, with no differentiation moat. #AI Agents#LLMs#Self-Hosted#Automation

    Open-source AI agents : web search, code sandbox, file analysis, agent teamwork, payments and sign-ups. Hosted or fully local.

    abcampbelTypeScript★89

  • C deepinfer-studio Visualizes the inference process in the browser without a GPU, solving the lack of an interactive environment for teaching how LLM inference works. Its 111 stars in 1.6 days come from AI education content spreading; monetization could be university course licensing or training platforms, but the audience is narrow and willingness to pay low; competitors are the many notebook tutorials and Transformer visualization projects. #Learning#LLMs#ML Research#Web Frontend

    Browser-based educational inference simulator and learning workbench. No GPU required. Real server integration and desktop app planned.

    duliangkuanPython★111

  • C dashboard-design-skill Distills styling lessons from 127 production dashboards into a Claude Code skill, solving ugly, state-less, inconsistent AI-generated admin UIs. It gains exposure from the skill distribution wave; commercialization is only a design-system subscription or component library, hard to make a standalone product; the risk is a flood of similar dashboard skills and replacement by native IDE templates. #Agent Skills#AI#UI Design

    A Claude Code skill for clean, honest dashboard UI: KPI tiles, charts, date pickers, ranked lists, states, layouts and a screenshot QA pass. Distilled from 127 shipped dashboards.

    abdiraimov-otabekLanguage unknown★97

  • C ian-xhs-carousel Lays out article notes as 3:4 Xiaohongshu image posts with 20 templates, solving creators' time-consuming layout work. Its 135 stars in 0.7 days make it a typical sample of the Claude Code Skill distribution dividend; monetization could be image-post generation SaaS or template subscriptions; the risk is Xiaohongshu's built-in layout tools and many similar layout sites, with the moat only in template aesthetics. #Agent Skills#Images#Content Writing

    Lay out articles, case studies, or notes as 3:4 Xiaohongshu image posts | 1080×1440 | 20 templates | light / dark | Claude Code Skill

    helloianneoHTML★135

  • C soloist-mpris Wraps Spotify's official Soloist player as an MPRIS service, solving the long-standing Linux problems of a bloated official client and broken media keys and playerctl. Its 89 stars in 1.9 days come from a real Linux desktop pain point; there is no paid scenario, at most distros packaging it as a default component; the risk is Spotify changing its protocol or blocking third-party wrappers, making it break anytime. #Linux#Audio#Streaming

    Spotify on Linux in ~60 MB: Spotify's official Soloist player as an MPRIS player, so bars, media keys and playerctl just work.

    TobiasVeiga00Rust★89

  • C php2shell-CVE-2026-109618-CVE-2026-109619-PoC Another PoC for the same CVE, offering a directly reproducible RCE exploit chain for penetration testers and incident responders; 571 stars in 0.6 days is disclosure-window squatting, not long-term value. There is no sustainable commercialization path beyond becoming a nuclei template, and the risk is the repo dying once the hype passes or being taken down by the platform. #Exploits#App Security#Network Security

    Proof-of-Concept (PoC) exploit for php2shell (CVE-2026-109618 / CVE-2026-109619), demonstrating remote code execution in vulnerable PHP environments. For security research and educational purposes onl…

    enessShell★571

  • C shrimpreel Packages a whole site that turns one photo into viral-template videos as a replicable scaffold, including login, credit packs, and fal.ai integration, solving the 0-to-1 build cost for AI video sites. It gains customers from AI video hype and the site-template business; monetization is selling templates or managed operations with a low ceiling; risks are severe template-site homogenization and payment/content compliance. #AI#Image Generation#Video#Marketing#Web Frontend

    Open-source template for "one photo → AI video that follows a viral template" SaaS sites, one site per trending keyword. Cloudflare Workers + D1/R2/KV/Queues, TanStack Start + Hono, Google sign-in, cr…

    qianshuixiaTypeScript★80

  • C simple-checker-v1 A Codex skill that breaks code into colorful flowcharts and verifies implementations against runtime evidence, solving the difficulty of validating AI-generated code. Its verifiability angle gains exposure in the skill wave; commercialization is only as a plugin to code-review SaaS, hard to charge for independently; the risk is Cursor/Copilot building the same capability natively and swallowing such skills. #AI#Agent Skills#Code Quality

    Codex skill: break down code feature modules and verify implementation with colorful flowcharts and runtime evidence — helps you write and update AI code.

    Monopoly0024Python★112

  • C douyin-to-obsidian Pulls Douyin videos, image posts, and collections into Obsidian with local ASR/OCR, fixing the problem that saved content can't be searched. It hit 81 stars in 1.4 days by riding both knowledge management and Claude Skill trends; monetization is limited to a personal subscription tool, with platform anti-scraping and copyright as risks and clipping plugins as competitors. #Knowledge Base#Agent Skills#Speech#Downloaders

    A portable /dy skill for cc-connect + Claude Code: Douyin videos, image posts and collections to Obsidian, with local ASR/OCR and private configuration.

    llm-xiaoyePython★81

  • C guizang-tv A clean Android launcher for Xiaomi TVs in China, addressing ad-heavy smart TVs and elderly users who can't navigate them. Its 141 stars in 1.4 days came from word of mouth in the TV-tinkering community, a real but limited need; monetization is hard beyond custom flashing services, with vendor updates blocking sideloading as the risk and mature launchers like Dangbei as competitors. #Android#UI Design#Productivity

    Guizang TV: a clean Android launcher adapted for China-market Xiaomi TVs, with large-text lists, wallpaper carousel, and direct HDMI access.

    op7418Java★141

  • C nano-deepinfer A CPU-first small inference engine prioritizing readability, with an honest description and restrained scope, suited to people who want to read source to learn inference; but nanoGPT-style teaching implementations already have several mature versions, and it's two days old with no performance or correctness comparisons. #ML Research#LLMs#Learning

    A small, readable inference engine inspired by DeepInfer. CPU first, built for learning.

    Alexander-QiuPython★13

  • C dragonwilds-inventory-planner A Windows inventory and equipment planner for RuneScape: Dragonwilds, filling a real need for a new game, plainly built. But companion tools like this are trivial to make, and it has no value outside that game's players. #Games#Windows#Productivity

    Dragonwilds Inventory Planner is a Windows companion project for RuneScape: Dragonwilds, focused on inventory organization and equipment planning. Explore its documentation and packaged release.

    MicroPheasantCSS★19

  • C claude-office Draws what Claude Code is doing as little people in an office in the browser. The idea is fun and demo-friendly, but it's essentially a visualization toy with limited help for actually using agents. #AI Agents#UI Design#Web Frontend

    A local tool that shows what Claude Code is doing as miniature office characters in your browser 🏢

    toplainsoftHTML★22

  • C HoloSplat-3DGS-Robot-Digital-Twin It connects 3D Gaussian Splatting scene registration and a ROS-driven mobile manipulator into a HoloLens digital twin, an interesting direction; but the same author has two repos on the same day, both lacking demos and hardware lists, looking more like internal lab modules than usable tools. #Robotics#3D#GPU

    3D Gaussian Splatting scene registration and ROS-driven mobile-manipulator digital twin modules for HoloLens spatial interaction.

    CyberyycC#★16

  • C LumaDock A macOS launcher for importing and starting Switch games, solving tedious emulator configuration; its 100 stars in 1.7 days come from handheld-emulation circles; it cannot be commercialized compliantly and can only rely on donations, and it directly invites Nintendo legal action; the risk is DMCA takedown and upstream nxnative changes breaking the project at any time. #Games#macOS#Piracy

    LumaDock — a native macOS launcher for player-imported Switch games, powered by nxnative

    Xc1YmSwift★100

  • C PS5_MarioGalaxy Statically recompiles Super Mario Galaxy and runs it on PS5 through a Win32 compatibility layer, via D3D12, vkd3d and RADV. The technical depth is real, but it is a one-off experiment that ordinary developers cannot install or use; its reference value lies mainly in the recompilation pipeline. #Game Porting#Reverse Engineering#GPU#ps5

    Super Mario Galaxy on the PlayStation 5: Nebula's static recompilation on a Win32 compatibility layer, Direct3D 12 on vkd3d and RADV

    grichesC++★14

  • C luci-theme-ctmodem Gives OpenWrt's LuCI interface the tacky skin of an ISP optical modem. It is an amusing joke theme, cleanly implemented, but has almost no practical use beyond fun screenshots, appealing only to people who tinker with router looks. #UI Design#Linux#openwrt

    CT modem (home gateway/ONT) style OpenWrt LuCI theme

    hakureiyuyukoCSS★24

  • C antifouling-coating-simulator Turns antifouling coating formulas into a runnable simulator, serving chemical formula R&D rather than developers. Its 90 stars in 1.5 days look more like a novelty project, lacking signs of real industrial adoption; commercialization needs experimental data and industry endorsement, high barriers but high deal sizes; the risk is the author's own note that simulation is not real testing, and industry won't buy without measured data. #Science#Web Frontend#Data Analysis

    Antifouling coating formulation simulator (Android + iOS + Electron). Simulation ≠ measurement. Track A/C example data in public repo.

    honghetingHTML★90

  • C METAL-GEAR-SOLID-2-Sons-of-Liberty-for-Android A static recompilation of MGS2 HD to Android ARM64, tuned for Adreno and Mali, requiring your own game files. It's technically impressive and clearly says bring your own files, but it's a one-off porting experiment that ordinary users can't install and that isn't really a general tool. #Game Porting#Reverse Engineering#Android#GPU

    Metal Gear Solid 2 HD (NVIDIA SHIELD) para Android ARM64 — recompilação estática ARM32→ARM64, Adreno e Mali, LSFG, menus em 7 idiomas. Traga sua própria cópia do jogo. / Static recompilation port, bri…

    luisxl15C★22

  • C scivolo A Scala interpreter and AOT compiler written in Rust. The topic is unusual, but one line of description can't support such a big claim, with no examples, compatibility scope, or progress notes, so treat it as an experiment for now. #Build#Performance#Testing

    Scala Interpreter / AOT Compiler written in Rust

    reactivecoreRust★27

  • C HoloMotion-Spatial-Robot-Teleoperation HoloLens head-pose alignment plus gesture-tracking quaternion teleoperation connected to ROS, real engineering effort; but it requires both a HoloLens and a robotic arm to verify, and the repo has no video or demo showing results, so most people can't get started. #Robotics#3D#Hardware

    Head-aligned spatial robot teleoperation with hand-pose tracking, quaternion motion commands, clutching, and ROS integration.

    CyberyycC#★13

  • C php2shell An operational intelligence repo from the Dutch national cybersecurity center on the php2shell vulnerability, giving defenders a unified source for detection and mitigation guidance. Official backing brought 109 stars in 0.6 days; as a government advisory it has no commercialization path but may become a de facto citation source; the risk is that content freezes once disclosure ends and it is not a product. #App Security#Network Security#Exploits

    Operational information regarding the 2026 php2shell vulnerabilities

    NCSC-NLPython★109

  • C McCosmos Builds a 3D cartoon food universe from McDonald's MCP nutrition data, with a visual encyclopedia and interstellar exploration. It is a fun tech demo, but beyond proving MCP can feed data to a 3D frontend, it is hard to say who gets daily value from it. #MCP#3D#Web Frontend

    McCosmos · McDonald's Star Map | A 3D cartoon food universe based on McDonald's MCP nutrition data, supporting real product images, nutrition encyclopedia and interstellar exploration.

    ayingQAQTypeScript★51

  • C omarchy-machine-mark Generates a custom little badge for the Omarchy desktop based on machine specs and runtime state, a cute desktop trinket. The scope is small and the description honest, but the value is basically aesthetic. #Linux#UI Design#Productivity

    Your Omarchy machine draws its own little mark from its specs and stats.

    erikrjohanssonQML★22

  • C CapRocket An open-source desktop editor positioned against CapCut, listing a timeline, offline Whisper subtitles, GPU effects, and no watermarks or accounts. The feature list is appealing, but a day-old Electron project probably doesn't run yet, so watch whether it delivers. #Video Editing#Open-Source Alternatives#AI#capcut

    Free, open-source desktop video editor in the style of CapCut: timeline editing, offline auto captions (Whisper), GPU animated effects & transitions, music/SFX & stock libraries, and an AI assistant. …

    medhachimJavaScript★20

  • C php2shell-poc Yet another PoC for the same PHP heap UAF, with a crash and a so-called RCE framework. A dozen equivalent implementations appeared this week, and this one adds no extra exploit-condition analysis or stability notes. #Exploits#App Security#Debugging

    CVE-2026-109618/109619 PHP2Shell — Heap UAF PoC (crash + RCE framework)

    Is4yevPython★22

  • C research-introduction-writing Turns experience in writing CS paper introductions into guidance text, claimed to be distilled from CCF-A and Q1 journal corpora. It is a content and prompt repository rather than a tool, and its effectiveness cannot be verified. #Content Writing#Learning#Prompting

    Analyze, draft, or revise Introductions for computer science papers — especially ML, computer vision, NLP, multimodal learning, information retrieval, and data mining. Its guidance is distilled from a…

    einstine336Language unknown★13

  • C home-hq A wall-mounted family dashboard with calendar, chores, shopping and Home Assistant, built on Rails 8, SQLite and Kamal. It is concretely done, but it is a personal-use project that is hard to migrate to other homes, so its reference value exceeds direct use. #Self-Hosted#Productivity#Databases

    Home HQ: a family dashboard for a wall-mounted photo frame and phones. Rails 8, Hotwire, SQLite, Kamal.

    joelgaffRuby★13

  • C cc-connect-dy-author A Skill that scrapes a Douyin creator's posts into Obsidian notes and interview questions. It's mostly for personal use, depends on page structures that will break, and there are already plenty of scraping-and-organizing tools like it. #Agent Skills#Knowledge Base#Content Writing#douyin

    A standalone /dy-author Skill: organizes a Douyin author's works into Obsidian knowledge notes and interview questions, supports personal configuration and cc-connect.

    llm-xiaoyePython★56

  • C BaiduPCS-Swift Ferry, a native macOS Baidu Netdisk download client written in Swift. Attractive to people fed up with the official client, but the core of such third-party downloaders is bypassing speed limits, and the author doesn't state compliance boundaries; login stability and resume support are also unknown. #Downloaders#macOS#Networking

    Ferry — a native macOS Baidu Netdisk download tool. Native macOS client for Baidu Netdisk.

    yiquncodeSwift★14

  • C app-lens A local analysis toolbox plus MCP billed as 'making software internals visible', with a Windows installer. But the description stops at slogans and lists no concrete analysis capabilities, so it is impossible to tell how it differs from ordinary reverse-engineering or debugging tools. #Windows#MCP#Debugging

    AppLens software insight: make software internals visible. A local software analysis toolbox and MCP service, providing Windows installation and acceptance.

    JichaoChen1123TypeScript★14

  • C marco-familiar-dashboard A Spanish-language family wall panel built on Planify-Dashboard, wired to iCal, chore scheduling, shopping, and Home Assistant. It's a personal fork, so you're better off looking at the original project if you actually want to use it. #Self-Hosted#Productivity#home-assistant

    Marco de fotos convertido en panel familiar de pared: calendario iCal, turnos, tareas de los niños, compra, luz y Home Assistant. Basado en Planify-Dashboard de eusonlito.

    leontdiKotlin★46

  • C telegram-black-hole-delete A UIKit plus Metal demo that turns Telegram iOS's "clear history" into a Dynamic Island black-hole animation. The interaction demo is nicely done, but it's just a feature suggestion that won't make it into the official client, so its practical value is low. #iOS#Animation#UI Design#telegram

    Telegram iOS feature suggestion: Clear History as a black hole from the Dynamic Island. UIKit + Metal, ready to integrate.

    riveauxmiaSwift★12

  • C lightning A fast native Windows search tool written in Rust. Everything already nearly monopolizes Windows search, this one doesn't explain its indexing approach or how it differs, and 19 stars show no momentum. #CLI#Windows#Performance

    Rust-based native fast search tool for Windows

    sashpawar11Rust★19

  • C torchlight-rekindled Apparently a port of Torchlight to modern handheld devices, a direction that appeals to retro gamers. But the description is one line, with no target platforms, build instructions, or asset handling, so completeness cannot be judged. #Game Porting#Games#Build

    Torchlight build for modern portable systems

    GreenSwampDwellerC++★16

  • C Ornith-1.5-122B-TensorFold A recipe for running a 122B EXL3 quantized model with TensorFold on a single DGX Spark, practical for anyone who owns the machine. But it is essentially a set of run scripts, and the model name cannot be verified externally. #LLMs#GPU#ML Research

    Ornith-1.5-122B (EXL3-EAQ) on one DGX Spark with TensorFold

    MiaAI-LabShell★12

  • C Runflare-Proxy A relay service running on Runflare with a password-protected web panel. The description is one sentence and does not say what it proxies, who it is for, or how it differs from similar panels; 35 stars in a day looks more like riding platform hype. #Networking#Self-Hosted#Privacy

    a relay on Runflare with a password-protected web panel

    ThisIsDaraJavaScript★35

  • C deskstudio "Open a design studio inside the coding agent you already use"—beyond that slogan there is nothing verifiable: which agent is supported and what files it produces are unstated. 41 stars in a day looks more like attention from the title. #UI Design#AI Agents#Web Frontend

    Your own design studio inside the coding agent you already use.

    adityareddy-devHTML★41

  • C design-style A daily design-style motion short with a production guide and open source. It is essentially a content collection, good for learning styles from the visuals, but offers no reusable tool or library and is of limited practical help to developers. #UI Design#Video#Animation#Learning

    Design Style · one design style per day — motion shorts, style images and production guides, all source code open.

    guanmo-aiTypeScript★48

  • C BO2-Requiem-NX A fan port of Black Ops 2 to Switch via the Requiem project, technically worth a look. But there are no build instructions and no stance on asset files, and fan ports of commercial games like this usually do not last long. #Game Porting#Games#Reverse Engineering

    Nintendo Switch port of Black Ops 2 (Requiem Project)

    HDErickLanguage unknown★20

  • C nxnative-romfs-archive It splits the Switch RomFS into about a hundred files to install large games on iOS, a small technical trick. But its main use is moving commercial game data, which does not hold up legally. #Piracy#iOS#Games#Reverse Engineering

    nxnative + RomFS archive: pack a Switch RomFS into ~100 files so large games can be installed on iOS

    LNPANDYC★23

  • C Listing-Clock Notifies you when an item expires at a Minecraft server auction house, a tiny notifier for players of one specific server. The audience is very narrow, the implementation simple, and it has no general value. #Games#Automation#Monitoring

    Notifys you when your items on donutsmp's auction house have expired.

    dropout12712Java★15

  • C mcd-group-order No description; the name suggests a McDonald's group-order script, and 86 stars in a day looks more like social spread than technical recognition. Without documentation, implementation quality and compliance cannot be assessed, so it scores low under the rules. #E-commerce#Automation#CLI

    gaojimmy27Python★86

  • C design-skills A collection of Agent Skills for AI image generation and visual design. It's a skills-and-prompts content repo lacking verifiable results, and several similar ones appear in this space every day. #Agent Skills#Image Generation#UI Design

    Design Skills: Open-source Agent Skills for AI image creation and visual design.

    kangarookingPython★13

  • C BuckshotRoulettePortable-BSHT A native C++ port of the paid game Buckshot Roulette to PS Vita, with only a one-line, unclear disclaimer and no word on asset sources or installation. It's a niche treat for the Vita homebrew scene, but asset legality and usability are unaddressed, making it look like a half-finished project. #Game Porting#Games#Hardware

    A native C++ port(able) of Buckshot Roulette. General Release of Liability on the Vita

    voidchecksamC++★12

  • C RCE-POCKET-POC-for-PHP2SHELL A 'pocket edition' PoC and experiment materials for the same PHP vulnerability, even more bare-bones than the neighboring Docker lab. With a dozen similar implementations already out, it adds no extra information. #Exploits#App Security#Network Security

    Pocket proof-of-concept and lab materials for PHP2SHELL remote code execution research.

    KarimUD7Python★19

  • C supermimic_vibed A SuperMimic reproduction vibe-coded in an hour, claiming LAFAN training in 1.5 GPUs. It's a reference implementation for motion tracking people, but the description is colloquial with no experimental comparison, making it more a personal practice log than reusable code. #ML Research#Animation#3D

    vibe coded supermimic within 1h, LAFAN passed within 1.5 gpuh training

    zita-chPython★20

  • C dsh-prompt-layer It adds a configurable system prompt to DeepSeek Harness, inserted after the built-in sections; meaningful only to users of that one harness, it's essentially a thin configuration layer with no evaluable technical content. #Prompting#Plugins#LLMs

    Configurable system-prompt section for DeepSeek Harness, rendered after every built-in section.

    AtinyFurinaJavaScript★15

  • C im-agent Described only as a vision-model-based AI operations agent, it never says which platform it operates, what the inputs and outputs are, or how to connect it. It is more vision than implementation, and cannot be installed or used today. #AI Agents#Multimodal#Marketing

    An AI operations agent based on vision models

    caixukun-jiTypeScript★16

  • C MutantPilot A controller that automatically operates Fargo's Mutant in tModLoader single-player. It serves only a small slice of Terraria players, is small in code and not general-purpose, so people in that circle can just grab it themselves. #Games#Automation#terraria

    Single-player tModLoader controller for Fargo's Mutant.

    snsididjdjsjsjC#★12

  • C ai-call-screen-share Self-hosted AI voice calls with screen sharing and camera, titled "Android shell edition," so it looks more like companion code for a tutorial series than a runnable product. Useful as a reference for on-device AI calling, but the repo gives no deployment steps, so its value depends on follow-up articles. #Speech#Multimodal#Android#AI

    See yourself on the call: voice calls, screen sharing and camera for a self-hosted AI (Android wrapper edition)

    wave2234HTML★14

  • C Officely Four AI robots chat and collaborate in a 3D virtual office, with a feature list running from calendars to a Gmail plugin—a classic vision-first project. There is no runnable evidence and no starting point, making it more a concept demo than something usable. #AI Agents#3D#Collaboration

    A virtual agentic office in 3D where four AI robot agents (Evan, Lummi, Athena, Zed) chat, collaborate and get work done. Zed, the reception manager, takes your commands and delegates tasks. Includes …

    Tahirpathan-AiLabHTML★14

  • C ghostldr A tool for loading PS4 game backups on the PS5 via SWRR. Even though it states it contains no game files, its core purpose is loading pirated backups, so by the rules it gets a low score directly. #Piracy#Games#ps5

    Run a PS4 game backup inside SWRR on PS5 through Luac0re, without a kernel exploit (tooling only, no game files)

    asylumexpC★15

  • C awesome-generalizable-world-models A paper list on world model generalization and interpretability. The curation has value, but it is still an awesome-list content repo and not something usable. #ML Research#Learning#Science

    Research papers on the mechanisms, generalization, and interpretability of learned world models.

    jidaivitaPython★15

  • C openloc Calls itself an open-source TMS built for AI. A translation management system should explain its scenarios and integration, but the description is a single positioning line, so it's unclear how it differs from tools like Crowdin. #LLMs#APIs#Documents

    The open TMS built for AI

    modelfrontGo★18

  • C hermes-plugin-gameforge This plugin claims to handle sound effects, music, NPC voiceover, pixel art, 3D models, four engine integrations, auto-playtesting and one-click export — a classic do-everything project. Nothing in it suggests more than a thin wrapper over existing APIs. #Game Dev#AI Agents#Plugins#Image Generation

    Game development plugin for Hermes Agent: sound effects, music, NPC voices, pixel art, 3D models, Godot/Unity/Blender/Unreal, a playtester that actually plays the build, and one-click export.

    DECRUX9812Python★32

  • C ai-roadmap-2026 Yet another AI learning roadmap, split into five layers (fundamentals, LLM engineering, agents, evaluation, production) with free resources and practice projects. However detailed, it is still a list-style content repo, and 73 stars in half a day mostly reflects inertia around the roadmap genre. #Learning#LLMs#AI Agents

    AI Roadmap 2026: a beginner's path through Foundations, LLM Engineering, Agents, Evaluation, and Production, with verified free resources and a mini-project per layer

    h9-tecLanguage unknown★73

  • C eve Yet another always-on personal assistant: web chat, Telegram, proactive reminders, long-term memory, app integrations, plus a builder for one-click deployable copies. The scope is enormous, and nothing shows where it beats existing assistant frameworks. #AI Agents#LLMs#Self-Hosted

    Always-on personal AI assistant with web chat, Telegram, proactive reminders, long-term memory, and app integrations. Includes a builder that deploys configured copies in one click.

    EnergyBoyTWTypeScript★16

  • C Plumb 'An agent harness for inverse physics in dynamic scenes' is a vision statement, with no indication of inputs, outputs, or example runs. Nobody can currently tell whether it does simulation, reconstruction, or something else. #AI Agents#3D#ML Research

    An Agent Harness for Inverse Physics of Dynamic Scenes

    Inception3DPython★15

  • C usf2fc-to-cps2-compile No description; the name points to some format-to-CPS2 compilation pipeline, but inputs, outputs, purpose, and how to run it are all unstated. Usability cannot be judged, nor whether it repackages someone else's tool. #Games#Game Porting#Reverse Engineering

    Zer0-NexuSLanguage unknown★15

  • C php2shell-CVE-2026-109618-109619 A second php2shell repo with the same name prefix, with a description adding an ASLR chain and patch analysis, but it appeared almost simultaneously with the previous one, suggesting a mirror or name grab. The core CVE ID can't be verified and PoC quality can't be judged, so wait for official records. #Exploits#App Security#Network Security

    CVE-2026-109618/109619 — PHP session.upload_progress pre-auth RCE. Exploit research: remote ASLR chain, PoC & patch analysis.

    xuehuolllPython★13

  • C Aether-EOS-R 'A robot runtime connecting hardware, perception and skills' is an architecture manifesto, with no supported hardware list, no examples and no install method. It cannot be installed or run today. #Robotics#Embedded#SDKs

    The robot runtime for IslandOS Aether, connecting hardware, perception, and skills through shared interfaces for robot state and actions.

    Island-TeamC++★18

  • C api-mesa Bundles 150+ pay-per-call APIs behind a single key and playground, essentially an API resale catalog with no technical increment and a description full of marketing terms. #APIs#Marketing#Data Analysis

    APIMesa — 150+ pay-per-call APIs for AI, web scraping, data, SEO, lead generation, media & developer utilities. One API key, live playground.

    cporter202Language unknown★12

  • C railway-anon-xray An undescribed xray relay deployment script whose purpose can be guessed, but with no usage instructions, configuration options or risk warnings. Repos like this usually have a very short lifespan. #Networking#Self-Hosted#Privacy

    free-dam-manPython★16

DSkip63
  • D .github A desktop tool for installing and switching DLSS 5 versions, saving players the hassle of manually swapping DLLs per game; 86 stars in one day fits the short burst typical of game-tool repos but is also suspicious. Mature open-source DLSS Swapper already owns this niche, leaving little room to charge; the risk is NVIDIA signature checks and driver blocking. #GPU#Windows#Downloaders

    DLSS 5 Swapper brings DLSS 5 installation and configuration into one desktop application, so you don't have to manage every game modification by hand. It can detect games from supported libraries, wor…

    DLSS-5-Swapper-Download-2026Language unknown★86

  • D ai-jailbreaks It compiles 35 paste-ready jailbreak prompts covering 25+ models, serving red-team and security evaluation scenarios; 206 stars in 0.4 days rides the natural spread of model-safety topics; there's no monetization path, at most it becomes an evaluation dataset; risks are vendors patching fast and platform bans, with competitors being existing open-source red-team prompt sets. #Jailbreak#LLMs#Prompting#AI

    AI jailbreaks: 35 paste-ready AI jailbreak prompts for 25+ frontier and open-weight models - system prompts, AGENTS.md specs, and Ollama Modelfiles. Built for AI security research and red-team testing…

    Alejandros17487Language unknown★206

  • D Sumi.css A minimalist classless CSS framework from the same source as HN, solving zero-class quick layout for small projects. Its 114 stars mainly come from 71 HN points, monetizing the author's existing influence; CSS frameworks have no paid scenario and rely on sponsors; the risk is that water.css and pico.css hold the space long-term, making it hard for a new framework to become a default dependency. #Web Frontend#UI Design#Performance

    A minimalist, classless-first CSS framework.

    vladocarHTML★114

  • D chat-keeper A 1:1 clone of the Claude web UI with a direct-connect mirror for mainland China, solving access restrictions rather than a technical problem; 135 stars in 1.9 days came from free availability, not product strength. It cannot be commercialized compliantly, account and quota sources are unsustainable, and the risks are high: vendor bans, key leaks and repo takedowns, so it is not worth following. #AI#LLMs#Crack

    chat-keeper focuses on providing Claude mirror sites with a 1:1 replica of the official site experience, directly accessible from mainland China.

    h563kCSS★135

  • D mcd-gacha-collector A card-collecting helper built on the McDonald's China MCP, a hackathon project solving the collection mechanic in a marketing campaign. Its 92 stars in 0.9 days come from MCP buzz and brand novelty; there is no commercialization room, it relies on non-public APIs, and it expires when the campaign ends; the risk is the API closing anytime and a cease-and-desist from the brand, so it is just a fun demo. #MCP#Agent Skills#Games#E-commerce

    McCard Collector · A card-collecting Skill + web version based on McDonald's China MCP (entry for the McDonald's Programmer Creative Development Competition)

    ilAoCPython★92

  • D arn0-tv A custom launcher written for the author's own Samsung Q70B (Tizen 6.5), essentially a personal device configuration. Nothing says whether it adapts to other models, so others basically cannot use it. #UI Design#Web Frontend#tizen

    A custom home screen for a Samsung Q70B (Tizen 6.5)

    NelsonEUJavaScript★22

  • D linux-taskmanager One sentence, "a task manager for Linux," with no screenshots and no explanation of how it differs from system monitors or btop. The category is long mature, and this project offers no differentiation. #Linux#Monitoring#Productivity

    Taskmanager For Linux

    anuzdhkRust★12

  • D product-trading A PHP virtual goods trading site selling accounts, skins, and game cards. Functionally it's just another generic storefront, with no explanation of how "secure trading" is delivered; account trading itself sits in a gray area, and code quality and hosting responsibility can't be judged. #E-commerce#Games#Web Frontend

    A virtual goods trading platform for accounts, skins, items and game-time cards, with secure transactions

    zs954PHP★52

  • D Discord-profile-template A Discord-style personal homepage template with animations and social links. Free alternatives abound online, and the description stays at generic selling points with nothing distinctive. #Web Frontend#UI Design#Social

    A customizable, open-source Discord-style bio website template with smooth animations, social links, and a clean design. Live demo available.

    Nero298JavaScript★13

  • D rec "Our lightning-fast inference engine," with nothing else: no supported models, no benchmarks, no usage. It's a marketing one-liner, so there's no way to tell how it differs from vLLM or SGLang. #LLMs#Performance#GPU

    Our lightning-fast inference engine.

    recursive-ossPython★52

  • D clove-api-1.0.6 The title only says source code for some API 1.0.6, with no clear problem being solved; it looks more like a mirror or leaked source of a paid API. Its 90 stars in 0.6 days come from spreading free access to a paid API, not product strength; there's no commercialization path and infringement risk; evidence is severely lacking, so not worth following up. #APIs#Reverse Engineering#Networking

    Clove API 1.0.6 source code

    xue-y-suGo★90

  • D couch-pip One line, "PiP for Steam Machine," with a typo and "Coded with Claude" in the description; it's unclear whether it's a picture-in-picture plugin on Steam or a standalone window, or whether it only fits the author's own device. #Games#Video#Windows

    PiP for Steam Machine - Coded with Calude

    CaraJean206TypeScript★17

  • D dsh-moe-lite A lightweight theme skin for DeepSeek Harness, purely cosmetic and serving only one harness. Beyond screenshots there is nothing to evaluate. #UI Design#Plugins#LLMs

    Lightweight Whale Musume theme for DeepSeek Harness.

    KurosawaGeekerJavaScript★18

  • D php2shell-nuclei-templates The README has nothing but a name, so it can only be guessed to be a nuclei template for the same CVE; it states neither detection logic nor false-positive rate, making it indistinguishable from the pile of same-named templates this week. #Exploits#App Security#Network Security

    AdemkingPowerShell★12

  • D .github A CS2 skin tool that injects to bypass the paid skin system, essentially a cheat repo funneling traffic; 244 stars in 1.4 days likely means star farming toward paid downloads. There is no compliant commercialization path, only gray distribution, with risks of Valve bans and GitHub copyright complaints, so skip it. #Game Cheats#Games#Crack

    🔥 cs-2-cs2-skin-changer-skin-swapper-skinchanger-skinswapper-all-skins

    cs-tool-sHTML★244

  • D YouTube-AI-Video-Editing-System Bundles a teleprompter, editing agent, JSON, Claude Design, and motion graphics into a "YouTube AI editing system." The buzzwords are piled high, but it doesn't say which parts are runnable code, a typical concept repo. #Video Editing#AI Agents#Content Writing

    Open-source YouTube AI video editing system: Eyeline teleprompter, editing agents, JSON, Claude Design, animation and motion graphics.

    sirioberatiHTML★14

  • D .github A Valorant skin changer that serves the demand for skins without paying, a typical cheat/crack repo. Its 242 stars in 1.9 days fit the usual star-farming pattern for such repos rather than real traction; the only monetization is bundle downloads and gray channels, so it can never become a standard dependency; risks are Riot anti-cheat bans and DMCA complaints. #Game Cheats#Piracy#valorant

    🔥 valorant-weapon-skin-changer-swapper-all-skins

    val-toolsHTML★242

  • D dsh-bili-humanism A prompt layer that adds a 'personality' to DSH and writes reflections into human-world.md. It reads as prompt content rather than mechanism, with most of the description devoted to philosophy, so it is unclear how behavior actually changes once installed. #Prompting#LLMs#Plugins

    Train DSH's persona with the colorful human world: by default accumulates letter reflections in a dedicated human-world.md; allows connection with external plugins via DSH. History, favorites, likes a…

    InariNGT-OfiicialJavaScript★12

  • D ACNHRecomp No README; from the name it is a static recompilation project for Animal Crossing: New Horizons. Build method, asset handling and completeness are all blank, so it cannot be assessed. #Game Porting#Games#Reverse Engineering

    NycolazsC++★18

  • D CVE-2026-109618-poc An empty-description repository that draws clicks only through a CVE number, with no reproduction steps or environment requirements; seen alongside another same-number repo from the same period, it looks like ID squatting. #Exploits#App Security#php

    enactive-systemsPython★26

  • D php2shellbt A Python scanner for the php2shell CVE series that emphasizes having no external dependencies. It does one thing and is nearly indistinguishable from the several same-topic scanners that appeared the same day. #Exploits#App Security

    Python scanner for CVE-2026-109618, CVE-2026-109619 and the php2shell session chains (handler mismatch + LFI). No external dependencies.

    pentestershellPython★12

  • D .github A Roblox script executor with a cheat menu serving in-game cheating, with very little technical depth. Its 240 stars in 1.4 days are highly isomorphic with other star-farmed repos and untrustworthy; there's no compliant commercialization beyond ads and bundled installs, and the risks are Roblox legal action and suspected malware distribution, so skip it. #Game Cheats#Games#Scams#Windows

    🔥 roblox-scripts-executor-mod-menu-with-cheats-and-hacks

    rbl-toolsHTML★240

  • D FPV_DVR_CLEANER There's no README, and the name suggests a tool for cleaning FPV goggle DVR footage; it says neither what the input and output are nor how to use it, so it's impossible to judge whether it's worth trying. #Video#CLI#Windows

    BratwurstexeRust★12

  • D Showcase-Project-IPII A work showcase page with no description; it is unclear whether this is coursework, practice, or a real tool. With no stated purpose, it cannot be evaluated. #Learning#Web Frontend#Documents

    alpinjolalaHTML★17

  • D side-fleet Zero description, only a language and a name, so it is unclear whether it is a scheduler, agent orchestration or something else, and usability cannot be judged. #Automation#CLI#SDKs

    foodakaRust★21

  • D .github A League of Legends skin changer that bypasses the paid skin system, a typical cheat funnel repo. Its 95 stars in 0.5 days match the structure of other star-farmed repos and don't reflect real demand; there's no compliant monetization, only bundled downloads, and the risk is Riot's anti-cheat detection and account bans, so skip it. #Game Cheats#Games#league-of-legends

    🔥 league-of-legends-lol-skin-swapper-skin-changer

    lol-tools-newHTML★95

  • D php2shell A PoC for CVE-2026-109618/109619 that only gives technical details and usage, with no background or impact scope. php2shell repos clustered on the same day and the source is hard to verify, so don't test it on a real machine. #Exploits#App Security

    PoC for CVE-2026-109618 / CVE-2026-109619 (php2shell): PHP session.upload_progress UAF. check = GET fingerprint + bulk JSONL; exploit = smart 618 (multi-path trigger) or 619 with --call. Stdlib only.

    murrezPython★18

  • D switchsplat No description; judging by the org and name, it may relate to 3D reconstruction or Gaussian splatting research. There's no README explaining methods, data, or results, so it can't be evaluated. #3D#ML Research#Images

    theialabPython★13

  • D ip-please-mobile Wraps ip-please for Android and iOS, with a description saying only "mobile version" and nothing about whether it checks public IP or something else, how to install it, or how it differs from the original. Not enough information to judge, and 14 stars prove little. #Networking#Android#iOS#Privacy

    Android & iOS version of ip-please

    zhcalvinJavaScript★14

  • D rtrvr-bulk-9router It mass-registers rtrvr.ai accounts, grabs Firebase tokens and session cookies, then pools the free quotas into an OpenAI-compatible pool—this is abuse of someone else's free tier, not a tool. It breaks as soon as authentication changes, and there's no legitimate use case. #Scams#Secrets#LLMs#APIs

    Bulk-create Retriever AI (rtrvr.ai) accounts, harvest Firebase tokens and session cookies, and serve 10 free managed models through an OpenAI-compatible pool wired into 9Router.

    0xgetzJavaScript★13

  • D resinfer Zero description; the name suggests something inference-related, but there is nothing verifiable: which models are supported, how to launch it, and who it is for are all unclear. Nobody can install and use it today. #LLMs#SDKs#APIs

    ShawRongPython★12

  • D switch-macos-launcher An undescribed macOS launcher with no screenshots or explanation, so it is impossible to tell how it differs from existing options like Spotlight and Alfred. #macOS#Productivity#Automation

    elisezhu123Swift★15

  • D Hertzon A zero-description repository whose name gives no clue what it does, with no install or usage hints, so its usability cannot be assessed. #LLMs#SDKs#APIs

    c8501024Python★29

  • D php2shell One of the php2shell CVE PoCs, with a one-line description and a CVSS score. Same-topic repos clustered the same day and the source is unclear, so don't run it on a real machine. #Exploits#App Security

    CVE-2026-109618/109619 PoC — PHP session.upload_progress heap use-after-free (CVSS 9.2)

    syrex1013Python★32

  • D glass-home-samsung-tv An undescribed JS project that, judging by the repo name, replaces the home screen on Samsung TVs. Several similar projects appeared the same day, and without documentation its feasibility and risks can't be assessed. #Web Frontend#UI Design#tizen

    victaphuJavaScript★51

  • D liberer-la-france It puts five simplified French legal proposals into a repo, essentially a policy essay rather than software, with nothing runnable or reusable; for a developer leaderboard it's not a candidate. #Documents#Social#Content Writing

    Cinq propositions de simplification du droit français par @alexmasmej (https://x.com/alexmasmej/status/2112088220804722917?s=46)

    AlexMasmejPython★12

  • D cuzinho-dev An undescribed TypeScript repository whose name offers no clue, so it is unclear whether it is a portfolio or a tool and cannot be assessed. #Web Frontend#Learning#Productivity

    eduumachTypeScript★32

  • D php2shell-checker An undescribed php2shell checker that nonetheless sits under the openai org name with 367 stars in a day, so both the naming and star count look abnormal. Don't download or run it until its origin is clear. #Exploits#App Security#Scams

    openaiJavaScript★367

  • D php2shell-poc Yet another standalone PoC for the php2shell vulnerability, giving only a CVSS score and a one-line summary. It's hard to distinguish from the batch of same-prefix repos that appeared the same day, so skip it. #Exploits#App Security

    Independent security research on the critical PHP session.upload_progress vulnerabilities (CVSS v4.0 9.2).

    scallydimaPython★15

  • D .github The repo is literally named .github, gained 76 stars in five hours with no code, and is a typical squat/mirror shell redirecting to some DLSS component replacement download. There is nothing here to evaluate. #Games#Windows#GPU

    DLSS 5 Swapper is a Windows utility designed to help users install and manage DLSS 5 neural-rendering components in compatible games and supported emulators.

    DLSS-5-Swapper-Update-DownloadLanguage unknown★76

  • D wii-tv An empty README, a Java project, and 74 stars in under a day, yet no text explaining what it does. The star velocity does not match the content at all. #Games#Reverse Engineering#wii

    MauroProtoJava★74

  • D interview-project-examples A retired Ostorlab interview project, kept for reference. The authors themselves mark it as not a product, so it has no value for readers of a ranking list. #Learning#App Security#Testing

    Retired example projects from Ostorlab's interview process, kept for reference.

    OstorlabPython★17

  • D TheDarkKnightRisesRevived No description and no language tag; only the name hints at a game connection, so it is impossible to tell what it does or whether it is a resource mirror or a name squat. Such a repository should not occupy a ranking slot. #Games#Game Porting#Reverse Engineering

    RedMemoriesLanguage unknown★19

  • D NEWSLETTER-PIZZA No explanation at all, and neither the name nor the language reveals what problem it solves. Usability cannot be judged and there is no way to start using it, so by the rules it gets no score. #Content Writing#Email#Marketing

    Shaady6ixPython★12

  • D Camtasia A piracy distribution repository named as a full Camtasia 2026 installer, not a tool project; its 80 stars in 0.3 days come from search traffic on crack keywords, unrelated to real adoption; no commercialization or standardization potential, and high probability of bundled trojans; skip it. #Crack#Piracy#Video Editing

    Camtasia 2026: Full working installer for professional video editing, screen recording, and VFX. Ready-to-use desktop edition with all...

    ThresholdSummonerLanguage unknown★80

  • D KMS-Pico A reskinned distribution repository for the KMS Pico activation bypass tool, serving piracy activation needs with no technical innovation; 80 stars in 0.3 days is typical star-farming for traffic, not a demand signal; no compliant commercialization path and clearly against platform policy; the risk is bundled malware and takedown, so skip it. #Activation Bypass#Crack#Piracy#Windows

    Kms Pico Free 2026: Full working installer for professional desktop software utility configuration, ready to run without activation hassles.

    TorrentthefoundryLanguage unknown★80

  • D CVE-2026-109618-exploit Exploit code for CVE-2026-109618 with no description. The php2shell topic spawned many identical repos in a day, and anonymous PoCs like this can't be verified, so running it is more risk than reward. #Exploits#App Security

    junairePython★82

  • D kitapokur Nothing but a Turkish name and 15 stars: no explanation of what it does, who it's for, or how to run it, so it's unclear whether it's a reader or a practice project. The repo is under a day old with no usage signs, so it's not worth opening until the author adds a README. #Documents#Productivity#Web Frontend

    ka1nshaJavaScript★15

  • D tv-makeover The description is just one line, "A better TV," with no indication of what it changes, what device it runs on, or how to use it. Half a repo name isn't enough to judge. #Embedded#Hardware#Automation

    A better TV

    paisanos-ioShell★25

  • D Cinema-4d Distributes cracked installers under the guise of a Cinema 4D 2026 configuration tool; it's not a real open-source project. Its 82 stars in 0.3 days are typical star-farming for piracy search traffic with no technical evidence; there's no compliant monetization and a high chance of bundled malware, so skip it and treat it as a risk signal. #Crack#Piracy#3D

    Professional Cinema 4D configuration utility for 2026, offering a full working installer with all essential features pre-configured and ready to run.

    shareshellreleaseLanguage unknown★82

  • D BorisFX Another cracked-distribution repo named after a full BorisFX 2026 installer, serving demand for pirated effects plugins. Its 80 stars in 0.3 days are bulk star-farming for traffic with no product value; there's no compliant monetization path and it violates platform policies, with risks of bundled malware and legal complaints, so skip it. #Crack#Piracy#Video

    Full working installer for Borisfx 2026, your professional VFX solution ready to run.

    slaterockwebLanguage unknown★80

  • D stacked-dither No description; from the repo name it looks like a small front-end project related to dithering effects. There's no explanation of purpose or usage, so it can't be evaluated. #Images#Web Frontend#Animation

    stackedlolTypeScript★12

  • D BootcampAI_26_27 A collection of course notebooks from an AI bootcamp's 2026/27 program, teaching material rather than a tool. Unless you are actually enrolled in that class, it has no practical value. #Learning#AI#Documents

    Kursy Bootcamp AI DeepDive 2026/2027

    DeepDive-UGJupyter Notebook★13

  • D gomax-bench-1 An HTML repo with no README body, so it is unclear what the benchmark measures, how to reproduce it, or where the results are. Even what the 'bench' in the project name refers to is unexplained. #Benchmarks#Performance#Testing

    GOMAX-LABHTML★18

  • D Steam-Data-PC-Widget No README; from the name it is presumably a small Steam data widget for Windows. What it does, who it is for, and how to install it are all missing. #Games#Windows#steam

    KaganAytenPython★12

  • D -Fixer-unexpected-error-1413677057-Wardogs The title is just "fix an unexpected error in some game," with no technical explanation—a typical error-code phishing/traffic-bait repository. Do not touch it. #Scams#Games#Windows

    Fixer unexpected error 1413677057 Wardogs - fix your game

    borderdelegateteachLanguage unknown★17

  • D php2shell No description at all, just another repo in the php2shell CVE PoC series. Seven or eight projects with the same name prefix appeared within a day from unknown sources, so running it is not advised. #Exploits#App Security

    iQingshanPython★21

  • D ZeroPulseTV A Java project with no description or docs; the repo name hints at TV or live streaming, but it's unclear what it actually does, so skip it. #Video#Streaming#Hardware

    steventa2024-lgtmJava★12

  • D csqtt-rebirth No description or explanation; the repo name suggests a remake or remaster of some game. It's impossible to tell whether this is a tool, assets, or a practice project, so skip it. #Games#Game Dev#Windows

    ttrroyyRust★26

  • D nexus-cli A Go CLI with an empty description, so it's unclear which nexus it connects to, what problem it solves, or who it's for. It got 19 stars a day after creation, but with no information to judge by, it isn't worth opening. #CLI#DevOps#Automation

    xamrayevGo★19

  • D LeakScope The description is a single mashed-together word, 'leaksdarkbreach', pointing at leaked data. It does not say what the tool does, and the content itself is a dataset that should not be circulated. #Piracy#Scams#Secrets

    leaksdarkbreach

    IvancastlCSS★22

  • D Mornye-Desktop Only one line saying it's coming soon to macOS, Windows, and Linux, with no indication of what the software is. All 35 stars come from the teaser, and there's nothing to evaluate. #macOS#Windows#Linux

    Mornye for macOS, Windows and Linux. Coming soon.

    mornyeappLanguage unknown★35

  • D Be-yourself The repo has no documentation at all, so it's impossible to tell whether it's a tool, notes, or a practice project; it falls into the unevaluable category. #Learning#Content Writing#Social

    suki052Language unknown★14

Show HN 2

S

Show HN: Omnara – Open-source alternative to Claude Managed Agents

Show HN

Turns the agent main loop into a durable harness of a Postgres event log plus stateless workers, solving real pains of losing state on restart, no multi-machine takeover, and idle cost burn. Its 2969 stars and founder's long post show a real user base; open-source core plus hosted version is the classic enterprise-standard play, directly replacing Claude Managed Agents; the biggest risk is Anthropic/OpenAI official hosted services crushing it.

Hey all, Kartik from Omnara ( https://omnara.com) here. We're building an open-source API for running durable agents. You create an agent with an API call, and Omnara runs the agent loop, gives the ag…

#AI Agents#LLMs#Self-Hosted#Open-Source Alternatives#claude

▲16

A

Show HN: Computer use on the phone itself: a 2B model that operates Android apps

Show HN

Runs a 2B small model on-device to drive Android app operations, solving GUI agents' latency, privacy, and API costs from cloud LLMs. Its 12 points are low but on-device computer-use is a real gap and easy to spread; paths are OEM preinstallation, accessibility automation, or mobile automation testing SDK licensing; the biggest risk is Google's own on-device agent and system permission lockdowns.

#AI Agents#Android#LLMs#Multimodal

▲12

Hot New Repos

  1. pengchujin/aerial-tv★2044 A clean launcher for China-market TVs: an open-source Android TV launcher designed after tvOS, no ads, auto-takes over on boot. A tvOS-style launcher for Xiaomi, TCL and other China-market TVs.
  2. ikow/php2shell-poc★921 CVE-2026-109618/109619 PHP session.upload_progress Exploit Research — POC & Analysis
  3. eagle-hq/EAGLE-4★719 Official Inference Code of EAGLE-4
  4. eness/php2shell-CVE-2026-109618-CVE-2026-109619-PoC★571 Proof-of-Concept (PoC) exploit for php2shell (CVE-2026-109618 / CVE-2026-109619), demonstrating remote code execution in vulnerable PHP environments. For security research and educational purposes only.
  5. dominikmartn/harp★531 Run many Claude Code and Codex agents side by side, in one sidebar that tells you which one needs you. An unofficial fork of Warp.
  6. NVIDIA/flare-local-inference★500
  7. 9bie/php2shell-bypass-aslr-exp★418
  8. openai/php2shell-checker★361
  9. renjiegod/dyt-ashare-cli★301 DYT CN: reproducible China-market research CLI and agent skills
  10. WoYun520/dsh-purge★283 DeepSeek Harness Purge (mirror of mmparko/dsh-purge) — full mirror: all branches / 100 tags / complete commit history. Original author credit in README.

Star gain 1d

  1. morluto/rea+4771★357652
  2. Gathin23/tv-vibehack+3807★19034
  3. Gathin23/cleanyourtv+3804★19037
  4. cloudflare/security-audit-skill+1421★36413
  5. Ebony-Vinyl/dsh-our-free-model+1387★33792
  6. zouyuxuan122/dsh-our-free-model+1378★33794
  7. deepseek-ai/DeepInfer+1333★17803
  8. DietrichGebert/ponytail+989★171588
  9. MadsLorentzen/ai-job-search+910★62923
  10. ikow/php2shell-poc+816★924

GitHub Trending · Today

GitHub's own order · heat relative to usual, not stars gained

  1. cloudflare/security-audit-skill+2012★36406
  2. vllm-project/vllm+441★94852 #LLMs#Performance#GPU#Self-Hosted
  3. stablyai/orca+1541★101211 #AI Agents#Productivity#Automation#codex
  4. headroomlabs-ai/headroom+474★77239 #LLMs#RAG#MCP#Performance#AI Agents
  5. DietrichGebert/ponytail+2079★171581 #Agent Skills#Prompting#AI Agents
  6. Comfy-Org/ComfyUI+939★141496 #Image Generation#AI#GPU#Plugins#stable-diffusion
  7. supabase/supabase+504★112984 #Databases#APIs#Self-Hosted#AI#postgres
  8. twentyhq/twenty+112★59146 #Open-Source Alternatives#AI#Productivity#salesforce
  9. shiyu-coder/Kronos+575★42812 #Fintech#ML Research#AI#Data Analysis
  10. akitaonrails/ai-memory+286★12313 #AI Agents#CLI#Knowledge Base#Productivity

GitHub Trending · This Week

GitHub's own order · heat relative to usual, not stars gained

  1. morluto/rea+129417★357612 #Reverse Engineering#AI Agents#Automation
  2. anthropics/claude-plugins-official+5703★43633 #Plugins#Agent Skills#AI Agents#CLI
  3. K-Dense-AI/scientific-agent-skills+4719★53576 #Agent Skills#AI Agents#Science#Data Analysis
  4. ruvnet/ruflo+3590★78210 #AI Agents#Automation#LLMs#Collaboration
  5. android/skills+1332★9147 #Agent Skills#Android#AI Agents#Learning
  6. NationalSecurityAgency/ghidra+5769★92623 #Reverse Engineering#Debugging#App Security
  7. debpalash/VoiceStudio+15125★80376 #Speech#Audio#Open-Source Alternatives#elevenlabs
  8. mksglu/context-mode+2968★31650 #AI Agents#MCP#LLMs
  9. heygen-com/hyperframes+5740★69197 #Video#Web Frontend#AI Agents
  10. cursor/plugins+2049★13395 #Plugins#Editors#AI Agents

GitHub Trending · This Month

GitHub's own order · heat relative to usual, not stars gained

  1. paperclipai/paperclip+24756★105673 #AI Agents#Productivity#Collaboration
  2. debpalash/VoiceStudio+46590★80376 #Speech#Audio#Open-Source Alternatives#elevenlabs
  3. vectorize-io/hindsight+28488★52362 #AI Agents#Knowledge Base#ML Research
  4. heygen-com/hyperframes+17177★69197 #Video#Web Frontend#AI Agents
  5. NVIDIA/OpenShell+8303★17009 #AI Agents#Privacy#Self-Hosted
  6. mksglu/context-mode+7915★31650 #AI Agents#MCP#LLMs
  7. cloudflare/security-audit-skill+16694★36406
  8. NationalSecurityAgency/ghidra+13471★92623 #Reverse Engineering#Debugging#App Security
  9. tashfeenahmed/freellmapi+11190★38951 #LLMs#APIs#Secrets#openai
  10. anthropics/knowledge-work-plugins+5802★31167 #Plugins#Productivity#Documents