tierlist.tech
中EN

09.18 Today's Picks

386 new projects found today, 220 rated.

Created 2026-09-16 – 2026-09-18Fetched 600Filtered 50Rated 40

A · Worth watching

A

node-webgl

RenaudRohlingerC

Brings ANGLE's Metal/D3D11/Vulkan backends into Node, enabling offscreen WebGL and three.js rendering without a browser, usable today for server-side screenshots and visualization output. It adds WebGL2 and multiple backends over headless-gl, at the cost of needing your own native build environment.

Headless WebGL 1 & 2 for Node.js, powered by ANGLE (Metal, D3D11, Vulkan). Render any WebGL or three.js scene without a browser.

#Web Frontend#GPU#3D#three.js

★32

A

techo5-dot

HuskerMinionPython

Flashes a second-gen Echo Dot with fully local firmware, removing Alexa, enabling all seven microphones, and connecting to Home Assistant as a voice satellite, with signed A/B updates and SSH. Hardware tinkerers will want to try it immediately, but it requires disassembly and flashing, and only covers this model.

Linux firmware for the Echo Dot (2nd gen): no Fire OS, no Alexa. A local Home Assistant voice satellite with all seven mics, Bluetooth speaker mode, signed A/B updates and SSH.

#IoT#Embedded#Self-Hosted#Speech#Linux

★38

A

idempotency-shield

agentscope-javaTypeScript

An Express middleware that caches responses by Idempotency-Key and replays duplicates, with memory or Redis storage, hitting the real pain of payment and order retries. Small in scope and accurately described, installable today; similar implementations exist, but this one works out of the box.

Express middleware that makes handlers safe to retry — replays the cached response for a duplicate Idempotency-Key instead of re-running it, with pluggable in-memory or Redis storage.

#APIs#DevOps#Databases

★17

A

dexterous-astra

jianglongyePython

A simulation environment, zero-shot scripts and trained checkpoints for dexterous-hand pen spinning and Rubik's cube solving, aimed at embodied AI researchers, and such reproducible in-hand manipulation resources are scarce. The limits are binding to a specific simulation stack and compute, making near-term deployment hard.

Dexterous pen spinning and Rubik's cube: sim environments, zero-shot scripts, trained checkpoints

#Robotics#ML Research#3D

★22

A

zeptun

NoisemuxZig

A Zig tun2socks engine delivering TCP/UDP/ICMP, SOCKS5, io_uring and a C API in one package, a convincing performance-oriented alternative to Go/Rust implementations. Proxy and traffic-forwarding developers can compile and try it, though it needs the Zig toolchain and lacks a ready client.

High-performance tun2socks engine in Zig with TCP, UDP, ICMP, SOCKS5, io_uring, and C API.

#Networking#Performance#Linux#CLI

★40

A

techo5-spot

HuskerMinionPython

Local firmware from the same author for the first-gen Echo Spot, keeping the round touchscreen, camera, Bluetooth and calling, also connected to Home Assistant. The engineering details are clear, but the approach matches the previous device and covers only one discontinued piece of hardware.

Linux firmware for the Echo Spot (1st gen): no Fire OS, no Alexa. A local Home Assistant voice satellite with a round touch face, camera, Bluetooth, phone calls and signed A/B updates.

#IoT#Embedded#Self-Hosted#Speech

★19

A

wesp-research

marcosd4hC++

Reverse-engineering documentation and tools for Microsoft's Windows Endpoint Security Platform, by a well-known figure in EDR. Directly useful to people building antivirus, EDR, or attack-defense research; the limits are that the platform is very new, the audience is narrow, and you must verify it yourself.

Technical documentation, reverse engineering research, and tools for the Windows Endpoint Security Platform (WESP).

#Reverse Engineering#App Security#Windows

★17

A

jevmlx

bnsd55Python

Constrained decoding with MLX on Apple Silicon, producing schema-conforming JSON in a single forward pass and skipping retries and post-processing. Practical for Mac developers running small local models needing structured output; nested schemas and streaming edge cases aren't yet addressed.

Jev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.

#LLMs#macOS#Performance#mlx

★24

A

sf4enhanced

fabeloperLanguage unknown

Adds rollback netcode to Steam's Super Street Fighter 4, a real need in the fighting game community, and similar changes have been proven on other old fighters. Currently the repo only hosts downloads with no source or implementation notes, so how far it goes depends on the author disclosing details.

Rollback netcode for Ultra Street Fighter IV (Steam). Downloads only.

#Games#Networking#Performance

★8

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

BHas a spark30
  • B skillbox This is a self-hosted, versioned agent skill library supporting MCP and client scopes, solving the lack of versioning, permissions, and auditing when teams share skills; the author is a well-known developer with built-in traffic. It could become the standard layer for enterprise skill distribution, but the risk is that the official skill repository is free and more deeply integrated. #AI Agents#Agent Skills#MCP#Self-Hosted

    Self-hosted, versioned skills library for AI agents. MCP, scoped clients, and optional Jev recommendations.

    kitzeTypeScript★82

  • B lead This is a Postgres full-text search extension for non-production environments, letting development, testing, and CI run TIN-compatible search SQL and solving the old mismatch between local and online behavior. Backed by PlanetScale, it could easily become a companion dependency alongside TIN; the risk is that it only works within their cloud product, locking in the audience. #Databases#Testing#CI

    A non-production Postgres text-search extension for exercising TIN-compatible application SQL in development, test, CI, and staging environments

    planetscaleRust★78

  • B tradingview-optimizer A Chrome extension from Roboquant that grid-searches TradingView strategies with walk-forward tests and heatmaps, sparing you custom backtest scripts, so retail traders can start directly. The feature scope is modestly stated, but watch extension permissions and the usual overfitting pitfalls. #Browser Extensions#Fintech#Data Analysis#tradingview

    Free, open-source TradingView strategy optimizer for Chrome. Grid search, walk-forward, heatmaps. By Roboquant.

    Roboquant-AITypeScript★28

  • B commit-miner From a familiar name in security tooling, it uses Jev to classify commit diffs and info, identifying fixes, CWEs and change types and generating risk summaries. Scope is clear with a CLI to try, but results depend on Jev and are hard to reproduce independently. #Code Quality#App Security#LLMs#CLI

    Classify Git commit diffs and messages with Jev. Bug fixes, security fixes/CWEs, and change types.

    devanshbathamRust★19

  • B mu300-linux Runs Linux (Ubuntu 26.04 with a self-compiled 5.4 kernel) on a ZTE F50/MU300 5G hotspot, with boot tools, kernel config and modem watchdog debugging notes. Hardware hobbyists will enjoy these first-hand notes, though the audience is tiny and the work is clearly ongoing. #Linux#Embedded#Hardware#Networking

    Linux (Ubuntu 26.04 + systemd, custom 5.4 kernel) on the ZTE F50 / MU300 5G hotspot (Unisoc T760 / UMS9620): boot tooling, kernel config, modem/PM watchdog bring-up, Wi-Fi, findings

    dikeckaanC★27

  • B ai-model-world A visual leaderboard of 556 large models as pixel characters, with data from Epoch, LiveBench, and Hugging Face synced hourly, static export with zero backend, and sources honestly stated. Its value is browsing and search experience, not a new evaluation method. #LLMs#Benchmarks#Web Frontend#Data Analysis

    AI large model world, a visualization site that anthropomorphizes 556 large models into pixel characters. Come in and see who is smartest right now, who writes code best, who is cheapest, and who just…

    liyupiTypeScript★30

  • B HFTENGINE A replay console for hftbacktest that visualizes the order book, queue position and matching latency, addressing unexplainable market-making backtests and hard-to-trace slippage. Quant teams pay well, so a commercial version is possible. Risk: it's essentially a wrapper around an open-source crate, and competitors are firms' in-house backtest visualizations. #Fintech#Crypto#Data Analysis

    Replay console for hftbacktest market-making backtests: order book, queue position of every resting order, feed and order latency, fills and the raw feed from the collector, on real recorded Binance F…

    mirkovicdevTypeScript★103

  • B github-ranking-audit It audits GitHub repository search ranking signals, helping maintainers diagnose why they cannot be found and which metadata to add; its 192 stars show exposure anxiety is widespread. It could extend into an open-source growth service or SaaS, but ranking rules are decided by GitHub's black box, the capability ceiling is low, and one redesign could render it obsolete. #Marketing#Productivity

    Audit your GitHub repository search ranking signals. Checks name, description, topics, README, stars, forks, and activity.

    nMaas8388Python★192

  • B openjev-sglang Wraps open-source models in a Jev-compatible interface via SGLang (prefill only) so local models join the Jev ecosystem; by ekzhang, so code credibility is relatively high. The approach is direct but covers only prefill, limiting capability and leaving compatibility to you. #LLMs#APIs#Self-Hosted

    Jev-compatible API endpoint based on open models (prefill-only)

    ekzhangPython★36

  • B ManiLoop A closed-loop evaluation framework for VLA/VLM robot policies, already integrated with MuJoCo, LIBERO and LeRobot, directly usable by researchers comparing policies. The value is a unified evaluation standard; the limit is that it's new, covers few benchmarks, and tasks must be added yourself. #Robotics#ML Research#Benchmarks#AI

    An open framework for closed-loop evaluation of VLA, VLM, and other robot policies across benchmarks, designed for fair and reproducible comparisons. Current integrations: MuJoCo tasks, LIBERO, GPT, a…

    Daniel-rmcPython★10

  • B surfacemap A reconnaissance tool for authorized testing that first organizes a target site's hosts, routes, parameters and trust boundaries into an evidence-backed handoff document before discussing exploitation, built on client-side JS analysis. Its positioning is restrained and audience clear; similar recon tools abound, and 7 stars means it's unproven. #App Security#Network Security#Automation

    Maps one authorized web origin's attack surface: hosts, technologies, routes, parameters, identities and trust boundaries, delivered as an evidenced, ranked hand-over document before a single payload …

    armourinfosecPython★7

  • B utuvo-orbit Small power, network, and audio toggles in the macOS menu bar, native SwiftUI, MIT licensed, bilingual Chinese-English, with a restrained scope and no hype in the description. Alternatives include Stats and OnlySwitch, so retaining users will depend on polish. #macOS#Productivity#Audio

    Your Mac. In one small orbit. A native macOS menu-bar app for power, network and audio. SwiftUI + AppKit · English / 繁體中文 · MIT.

    mickyyang-1407Swift★11

  • B jev-on-a-laptop Reproduces Jev-style parallel type decisions on Apple Silicon, with benchmarks and an HF demo, rare hands-on material for people wanting constrained decoding on local 1.5B-8B models. Its value depends on whether the reproduction is credible; currently there are only self-reported numbers and no comparison with other baselines. #LLMs#ML Research#macOS#Benchmarks#jev

    Unofficial study: Jev-style parallel typed decisions on stock 1.5B-8B models on an Apple Silicon laptop. Benchmarks, research notes, and a Hugging Face Space demo.

    rorshoppingPython★12

  • B bidcraft A conversational bid-document system using LangGraph to chain tender parsing, knowledge-base distillation and revision — a very concrete scenario that genuinely pains Chinese engineering bidders. Currently it's more an open-source shell of an internal product, lacking deployment docs and sample data; watch for now. #LLMs#AI Agents#RAG#Documents#langgraph

    BidCraft · conversational bid document authoring system: AI intelligently drafts technical bids, supports tender document parsing, lot selection, bid writing, knowledge base distillation, and conversa…

    zeronezerPython★11

  • B harness-bridge A bridge layer that points coding harnesses like Claude Code and Codex at any OpenAI/Anthropic-compatible endpoint, a real need. But it shipped a library, CLI, tray app and web UI in half a day, the scope is scattered, and it collides head-on with claude-code-router and LiteLLM. #AI Agents#LLMs#APIs#CLI#macOS

    Point any coding harness (Claude Code, Codex, OpenCode, Pi, OMP, Crush, Copilot, Grok…) at any OpenAI/Anthropic/Responses-compatible inference endpoint. Core library + CLI + macOS tray + web UI.

    0xSeroTypeScript★33

  • B gksdud A small macOS tool for quickly switching between Korean and English input methods, with a one-line purpose that Korean users would want. But Karabiner, Gureum and other alternatives are mature, and it doesn't say how it's faster or how to install. #macOS#Productivity#Automation

    A snappy macOS Korean-English input switcher

    codingnoyeSwift★24

  • B cdn-tunnel Forwards TCP/UDP traffic through a CDN using SOCKS5 + HTTP/2, with password authentication and an Android VPN wrapper, described concretely without hype. Suitable for self-built tunnels on restricted networks, but wstunnel and cloudflared are already mature, and nine days of stars do not show stability. #Networking#Privacy#Android

    TCP/UDP tunnel over a CDN (SOCKS5 + HTTP/2) with password auth, GET/POST/stream transports and an Android VPN wrapper

    pasteichJava★9

  • B fanduck Complete replication materials for the Microduck robot: live demo, bill of materials, 3D-printed parts and boards, so makers can build their own desktop robot pet. The value is in verified parts details; the limit is that it covers only this one piece of hardware and depends entirely on the author going forward. #Hardware#3D#Robotics#Embedded

    Fanduck · Microduck has been successfully replicated! Sharing live demos, bill of materials, 3D printing, boards, build tools, and a 3D mirror workbench. Build resources from our completed Microduck r…

    7757Language unknown★7

  • B fruitflysynth Turns a recently published neural dataset of Drosophila courtship song into an interactive synthesizer, combining science and playfulness, interesting to audio-synthesis or neural-encoding people. More fun than practical and small in code, it's an oddity worth a look. #Audio#Science#Biotech

    An interactive resynthesis of a male fruit fly's mating sounds utilizing the recently mapped neural dataset.

    bennjordanJavaScript★10

  • B oneplus-negroni-arch-linux A firmware repo porting Arch Linux to the OnePlus 10 Pro — mobile porting that few tackle head-on, of scarce value to those following phone mainlining. The limits are obvious: only one line of description, so flashing steps and usability are all on you to figure out. #Linux#Embedded#Hardware#oneplus-10-pro

    OnePlus 10 Pro arch linux firmware

    withsaltC★12

  • B UCAS-Desktop A UCAS campus toolkit: tray resident, lecture booking, course planning, online courses and PDFs, gathering scattered scripts into one Windows desktop app — highly practical for students. But it serves only one school, and automated course selection carries compliance risk. #Automation#Windows#Productivity#ucas

    UCAS Windows desktop assistant: runs in background in the system tray, Qingxin Classroom, lecture booking, course selection planning and automation, UCAS online video/PDF; modular integration with det…

    zhangzhendan-BerkeleyPython★9

  • B phicomm-s7-ha Replaces the Phicomm S7 body-fat scale's firmware with ESP8266, auto-connecting to Home Assistant via MQTT, so owners of this old scale can revive it. Scope is tiny and the writing solid, but it serves one discontinued device, so coverage is inevitably narrow. #IoT#Embedded#Self-Hosted#Health

    Phicomm S7 body fat scale ESP8266 replacement firmware — Home Assistant MQTT Discovery

    PraepesC++★8

  • B higgsfield-api-skill Wraps Higgsfield's image and video generation into a skill coding agents can call, with model docs and resumable tasks; convenient for existing users, but it's a third-party API wrapper replaceable by switching platforms. #Agent Skills#Image Generation#Video#APIs

    On-demand Higgsfield API image and video generation for coding agents, with model-specific docs and resumable jobs.

    cth9191Python★22

  • B ReactantNitro.jl A Lux training wrapper from EnzymeAD that compiles and runs a whole experiment in one pass with Reactant, cutting boilerplate. It comes from a trusted team with a restrained scope, but it has only 8 stars, thin docs, and is basically useless outside Lux. #ML Research#Performance#GPU#SDKs

    Reactant-first training for Lux models: declare the experiment, compile once, train without boilerplate.

    EnzymeADJulia★8

  • B remiqora Wants to cram text-to-music, stem separation, MIDI transcription, LoRA fine-tuning and a multitrack DAW into one local Vue interface, listing all that within half a day of creation, so actual completeness is doubtful. Music generation already has frontends like ACE-Step; see which piece it finishes first. #Audio#AI#ML Research

    Local AI music studio unifying ACE-Step 1.5 and YuE2-3B in one Vue interface — text-to-music generation, stem separation, MIDI transcription, and LoRA fine-tuning, with a built-in multitrack DAW.

    inikolaxVue★22

  • B EchoMind A local voice-memory assistant: real-time transcription plus RAG Q&A based on what you have said, with a clear privacy story. But transcription plus RAG is already a common combination, and beyond Rewind and Limitless it will have to win on engineering details. #Speech#RAG#Privacy#AI

    Local-first voice memory assistant — capture speech, transcribe live, and ask a private RAG chatbot grounded in what you've actually said. No cloud required.

    nirbhay41120003Python★25

  • B AnyDoor Xposed-based global mock location on Android, focused on hooking system_server and erasing mock flags to evade detection. Technically substantial, but its purpose is bypassing apps' location checks for attendance, social and games — a typical evasion tool, not recommended. #Android#Reverse Engineering#Privacy

    Anywhere Door · Xposed-based Android system-wide GPS spoofer with anti-detection, map picker/joystick/route simulation, optimized for Chinese networks | System-wide GPS spoofer for Android (Xposed/LSP…

    zhaoyuxiangyyds-labJava★22

  • B seijaku MengTo's Three.js Kyoto courtyard walkthrough offers procedural buildings, gardens and seasonal changes you can stroll through in the browser, well judged. But it's a piece rather than a tool; only the scene code is reusable, and ten minutes gets you the effect. #3D#Web Frontend#Animation#UI Design

    A continuous interactive walk through a Kyoto residence, with procedural Three.js architecture, gardens, seasons and responsive photography.

    MengToHTML★13

  • B llmwiki-operational A persistent knowledge system using LLMs to continuously maintain technical manuals and SOPs, targeting the problem of stale ops documentation. But it lacks getting-started steps and effectiveness evidence, and doesn't explain how it differs from wiki-plus-retrieval. #LLMs#Knowledge Base#Documents

    A persistent LLM-maintained knowledge system for technical manuals, SOPs, and operational documentation.

    alicasuuPython★24

  • B jev-ultrafast Nearly two thousand stars in a day, but the repo has no description, leaving name and buzzwords to guess its purpose; the stars come from the Jev topic effect rather than verifiable output, so wait for clear capabilities and usage. #LLMs#AI Agents#Performance

    browser-usePython★1984

COrdinary80
  • C aibuildai-llm-posttrain-agent It claims to use a recursive self-improving agent to automate LLM post-training and tuning, hitting the real pain of high post-training labor costs and, if true, sellable to model teams. But with 49 stars and grand claims backed by thin verifiable evidence, it looks more like concept packaging; reproducibility and effect evaluation are the biggest questions. #LLMs#AI Agents#ML Research

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

    aibuildai-incPython★49

  • C AirCard It lets users reskin Apple Wallet cards on iOS 18 without jailbreaking, addressing complaints that the wallet interface cannot be customized; 270 stars in one day show the demand is real. But it relies on private APIs, has no commercialization path, and could be blocked by Apple or rejected from listing at any time, making long-term risk extremely high. #iOS#UI Design

    Apple Wallet Card Skinner for iOS 18+ (No Jailbreak Required)

    Mak5erSwift★270

  • C HEXIS It compiles SKILL.md into a state machine, using transition guards to fix execution order and confining LLM reasoning within states, solving the uncertainty of agent skill execution paths; the direction has value. But at 69 stars it is small and directly overlaps mature orchestration frameworks like LangGraph, with an unclear commercialization path. #AI Agents#Agent Skills

    HEXIS. Compile SKILL.md into state machines: deterministic execution order via transition guards, LLM reasoning scoped inside states.

    Worldbuilder013Python★69

  • C go-responder It rewrites Responder in Go to perform LLMNR/NBT-NS/MDNS poisoning with built-in fake authentication services to capture NTLM hashes, deploying more easily cross-platform than the original, so red teams can use it directly. Its 46 stars are natural growth in the security community, and a commercial version is possible; the risk is the mature original ecosystem and the gray-area use case. #Network Security#App Security#Secrets

    Responder is a LLMNR, NBT-NS and MDNS poisoner, with built-in HTTP/SMB/MSSQL/FTP/LDAP rogue authentication server supporting NTLMv1/NTLMv2/LMv2, Extended Security NTLMSSP and Basic HTTP authentication…

    bl4ckarchGo★46

  • C GBC2DMG Converts supported GBC ROMs into a form the original Game Boy can run, starting with Crystal 1.0/1.1 and including version detection. It's a legitimate little tool for modders and retro fans, but it supports very few games and each addition needs individual work. #Games#Game Porting#pokemon

    Convert supported Game Boy Color ROMs for the original Game Boy. Crystal English 1.0 and 1.1 beta support, with built-in version identification.

    ItsJustBshawnPython★10

  • C calamansi-ui An animated component registry for Next.js and Tailwind, overlapping with magicui and aceternity in the shadcn ecosystem, with 'small' as its only selling point. Component libraries live on sustained output and taste, neither visible in a library only days old. #Web Frontend#UI Design#Animation

    A tiny, open-source animated React UI component registry for Next.js and Tailwind CSS.

    fujiDevvTypeScript★14

  • C testdrive-enhanced A rewrite of 1987's Test Drive in SDL3 with a clear technical approach, aimed at retro players; but it still needs the original assets to run, its audience is narrow, and it is unclear how it differs from existing remakes. #Games#Game Dev#Game Porting

    Test Drive (1987) on an SDL3 engine

    kylofonC★10

  • C typesafe-adblock A Chrome extension that uses a small model to judge whether DOM elements are ads; the author states it is a toy, not a real ad blocker, and BYOK with no backend, which is rare honesty. But its accuracy and performance are not enough to replace uBlock. #Browser Extensions#AI#Privacy

    🧹 Fun project: a Chrome extension that asks a tiny AI decision model (TypeSafe Jev) "is this DOM element an ad?" and pops it off the page. BYOK, no backend, not a real ad blocker.

    realZachiJavaScript★19

  • C github-launch-checklist Scores a repo against ten self-check items before release, a straightforward feature; but it launched in the same batch as the same organization's "search ranking" repo, looking more like part of an acquisition funnel, and the scripts checking README and license are only a few dozen lines to write yourself. #Code Quality#Marketing#Automation

    Audit a GitHub repository before launch - ten readiness checks scored 0-10: name, description, topics, README, license, demo, installation, contributing guide and issue templates. Free and open source…

    repoboost-hqPython★43

  • C tunneldock Aims to use the ChatGPT web client to drive local development in reverse — a clever angle pointing at the real pain that the web version has no local file access. But the description is one Chinese sentence, unclear whether it's an extension, bridge or automation script, and it doesn't mention account risk. #AI Agents#Automation#Browser Extensions#chatgpt

    Develop on your computer through the ChatGPT web page

    t59688TypeScript★18

  • C obsidian-mobile-companion Uses a GitHub repository to bring Obsidian-style Markdown sync and offline reading to Android, addressing the pain of accessing note vaults on mobile; its 57 stars show a sizable Obsidian user base. But official sync is a paid and more reliable feature, so unofficial solutions are hard to charge for and easily replaced. #Android#Knowledge Base#Self-Hosted#obsidian

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

    zhdmm35Kotlin★57

  • C madgen A tool that automatically generates MAD-style music videos, aimed at users who want to make remixes but can't edit, a niche subject with a concrete target. For now there's only one Japanese sentence of description, with no input/output or asset dependencies shown, so how much time it saves over manual editing can't be verified. #Video#Audio#Automation#Content Writing

    Open-source software that automatically generates 音MAD

    0266stPython★15

  • C youtube-agent-skill A YouTube operations suite built from eleven Claude skills, covering hook formulas, retention analysis and Shorts clipping. Creators can pick out useful pieces, but it's a bundle of prompts and workflows, and terms like "virality engine" are overly marketing. #AI#Agent Skills#Content Writing#Prompting#Video

    Eleven Claude skills that run a YouTube channel: scripts off 21 hook formulas with a scored hook, title and thumbnail linted as one pairing, an edit decision list from your transcript, a retention rea…

    JakeschincariolPython★21

  • C claude-plugins A Claude Code plugin marketplace repo whose value depends entirely on the quality of the plugins it lists, while the repo itself is just an index plus install instructions; several similar marketplace repos have appeared in the past month, so first check whether any plugins are actually used. #Plugins#AI Agents#LLMs

    Claude Code plugins by Lahfir. Marketplace: lahfir-plugins

    lahfirJavaScript★23

  • C Patchx-OpenDeskRobot A desktop companion robot with voice, expressions, and motion, an in-house team project with an honest description; but there is no hardware list or software boundary, so external reproduction is costly and for now it can only be viewed as a reference design. #Robotics#Hardware#Embedded#AI

    Patchx-OpenDeskRobot is an open-source desktop companion robot, built by the PatchX team for our own desks. It listens, talks, and reacts with expressions and motion.

    patchx2027Language unknown★9

  • C advocaat A type-safe data Q&A client based on Jev, overlapping in approach with text-to-SQL clients, with the core model's reliability unverified; reaching 41 stars in one day looks more like a single wave of concentrated promotion. #LLMs#Databases#APIs

    A small, type-safe client for asking AI questions about your data, powered by TypeSafe Jev.

    pithingsTypeScript★41

  • C tars Turns an iPhone into the face of an agent on a Mac, inspired by Interstellar, an interesting direction; but the repo does not explain the communication method, which agents are supported, or whether it is a concept or runnable, so for now it is just an idea showcase. #AI Agents#iOS#macOS

    Turns your iPhone into the face of your agents on Mac. Inspired by Interstellar.

    reycnSwift★9

  • C Agora Argues for using git as shared memory for multi-agent collaborative research, with an arXiv number — an idea somewhere between interesting and vague. The repo is mostly a webpage with no runnable collaboration framework in sight, and the paper number needs independent verification. #AI Agents#ML Research#Science#git

    Agora: Git as Shared Memory for Collective AutoResearch (arXiv:2609.18094)

    yifanzhang-proHTML★14

  • C gpt_sub_analysis A security analysis of ChatGPT iOS subscription and entitlement flows, answering whether authentication can be bypassed; its 199 stars show high interest in subscription bypass topics. But it is a research draft mid-disclosure with no productization or charging space, and the risk is platform pressure after publication plus difficulty reproducing the conclusions. #App Security#iOS#Reverse Engineering

    Security analysis of the ChatGPT iOS subscription / entitlement flow (internal draft, coordinated disclosure WIP)

    yynxxxxxLanguage unknown★199

  • C procedural-film It automatically generates 30-second vertical short videos from a topic, drawing both visuals and music in JS, solving batch short-video production; its 58 stars come from AI video generation hype. But the implementation is just procedural drawing plus scoring, far behind models like Sora and Veo, with a thin moat and easy coverage. #Video#Animation#Agent Skills

    Agent skill: turn a topic into a 30s vertical film drawn and scored entirely in JavaScript. Reference film: the life of a monarch butterfly.

    kuhnhomeuk-cellJavaScript★58

  • C agnes-manga-studio A local-first AI comic-drama production desk that relies on an Agnes-compatible API; the concept sounds complete, but it does not say how far storyboarding, voiceover, and compositing are actually implemented, and one-stop generation tools are already plentiful. #Image Generation#AI#Content Writing

    Local-first AI manga drama production studio powered by Agnes-compatible APIs.

    liobububuJavaScript★12

  • C Chromium-for-windows-2000 Brings Chromium to Windows 2000, a clear and sufficiently hard goal that appeals to retro-computing enthusiasts; but the repo has no build steps or patch list, so for now it is more a manifesto than a usable product. #Windows#Build#Performance

    Chromium for Windows 2000

    mjd7989Language unknown★18

  • C Hierarchos-Native Claims native Transformer training and PEFT fine-tuning in Rust + Vulkan, a rare approach worth comparing with burn and candle; but it claims a full pipeline one day after the repo was created, with no benchmarks or minimal example, so wait and see the implementation. #GPU#ML Research#LLMs

    A native rust-vulkan backend for training hierarchos and modern transformer model architectures, supporting PEFT finetuning, full training, and inference

    necat101Rust★14

  • C panelalpha-engine Lets AI agents take over a self-hosted deployment engine, from the PanelAlpha team; the direction is reasonable but the copy is promotional and does not clarify the agent's permission boundaries or failure rollback, the two most critical points in ops scenarios. #DevOps#Self-Hosted#AI Agents

    The easiest way to self-host projects on your own server. Connect your AI agent and let it deploy, maintain and manage multiple projects on any stack, while keeping your server organized and under con…

    panelalphaPHP★11

  • C MacTV It lets a Mac control TV volume and switch input sources directly via HDMI-CEC, solving the small pain of a fragmented remote once the TV becomes a monitor. Its 63 stars show the need is real but very narrow, it depends on specific hardware and cables, and there is almost no commercialization space, making it a personal utility. #macOS#Hardware

    MacTV · TV Remote — makes the TV better as a Mac display. Adjust TV volume, switch input sources, and operate menus on the Mac via HDMI-CEC.

    pengchujinSwift★63

  • C pureterm This is an SSH/SFTP client built on ssh2 and xterm.js, solving the split between remote files and terminal operations, but the space is extremely crowded, already covered by Termius, Warp, and built-in terminals. At 45 stars it is early-stage and lacks differentiation or a reason to pay. #Terminal#Networking#CLI

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

    puremixaiTypeScript★45

  • C ask-jev-skill A skill wrapper that lets agents like Hermes call typesafe Jev for typed judgments, a reasonable idea, but the content is only a thin adapter that cannot be evaluated apart from Jev itself. Unless you already use Jev, there is no reason to install it now. #AI Agents#Agent Skills#LLMs#jev#hermes

    Skill for Hermes, and other agents, to ask typesafe's jev

    shantanugoelPython★10

  • C xianyu-guanjia Multi-account auto-reply and order management for Xianyu, a complete Electron plus FastAPI desktop implementation with real appeal to second-hand e-commerce sellers. But multi-account automation itself skirts platform rules, and the repo does not explain risk control or compliance boundaries, so use at your own risk. #Automation#E-commerce#AI#xianyu

    Xianyu multi-account auto-reply and business management tool — Electron desktop + FastAPI backend. QR code login, keyword/AI replies, order shipping management.

    shudaxiagPython★8

  • C skillranker A Rust CLI that prioritizes agent skills using session context, with Claude Code hooks and structured output, and the engineering discipline shows. But it needs a TypeSafe API key to run, essentially a client for that service, so its standalone value is limited. #AI Agents#Agent Skills#CLI#LLMs

    Rust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks, structured JSON, abstention, and local feedback. Requires a T…

    DicklesworthstoneRust★15

  • C friday Adds persistent memory to coding agents, a real direction but a crowded field with mem0 and similar solutions; the copy is full of emoji and 'amnesia' metaphors but never explains storage format, retrieval or integration cost, so no differentiator is visible. #AI Agents#LLMs#Knowledge Base

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

    friday-memoryHTML★29

  • C block-blast-solver Solves a single-player puzzle game like Block Blast with exhaustive search plus a value network, a narrow goal at small scale, good for seeing how reinforcement learning applies to a concrete problem. But it is essentially a practice-level solver that does not explain the search/network tradeoff, and can be read in ten minutes. #Games#ML Research#AI

    Block Blast solver using RL (exhaustive search + value network)

    rshk941Python★8

  • C gemini-transcribe-mac A Mac desktop app for real-time voice transcription plus smart notes, using the Gemini API — useful for people taking meeting notes. But the model version in the description doesn't match, and the docs are Korean only, so you must read the code to confirm how it actually calls the API. #Speech#macOS#AI#Productivity#gemini

    A real-time AI voice transcription & smart notes desktop app for Mac based on Gemini 3.5 Transcribe

    thinkthank92Python★7

  • C AdaptiveRAG-Robustness Paper companion code studying adaptive RAG robustness under query paraphrase changes, useful reference for retrieval-evaluation researchers. But it offers no directly integrable components and doesn't state reproduction costs, so value for ordinary developers is limited. #RAG#ML Research#Benchmarks

    Official repository for "How You Ask Matters! Adaptive RAG Robustness to Query Variations".

    YunahJangPython★10

  • C omarchy-ring-light-plugin Adds a white screen border to the Omarchy desktop as fill light for video calls. The idea is straightforward and the code small, but hardware lights and existing software options are plentiful, so the audience is very narrow. #Linux#Video#Productivity

    A white screen border that lights your face during video calls, for the Omarchy desktop.

    bjarneoQML★11

  • C DJ-Fly An experimental art project making a fly into a DJ, a novel idea, but the repo explains neither inputs and outputs nor hardware dependencies, so practical value is essentially zero; treat it as a curiosity. #Audio#AI#Hardware

    DJ Fly — an experimental project where a fly becomes the DJ

    insomnia-vipPython★16

  • C pg-jev A Postgres extension for asking table data in natural language, equivalent in approach to text-to-SQL; it relies on the unverified Jev model and does not explain schema injection or read-only permission control, two points that cannot be vague in database scenarios. #Databases#LLMs#AI

    Ask your Postgres tables questions in plain language. A PostgreSQL extension powered by TypeSafe's Jev.

    realZachiShell★13

  • C typesafe-snake Uses Jev to make a typed choice every frame to auto-play Snake; the code is short and the boundaries clear, making it a decent model-capability demo. The limits are equally clear: it's a demo, not a tool, and teaches nothing beyond how Jev handles constrained decisions. #Games#LLMs#AI#jev

    Snake auto-played by TypeSafe's Jev model: one System One choice per tick, legal moves and facts generated in code

    sorryccTypeScript★8

  • C Xbox360-Native-Ports A collection of launchers for recompiling Xbox 360 games to Linux, explicitly stating it contains no game data and requires your own legitimate copies, an honest boundary. But "launchers only" means the core recompilation work isn't visible yet, and its value versus the Xenia route remains unproven. #Game Porting#Reverse Engineering#Linux#Games

    Xbox 360 titles recompiled to native Linux - launchers only. You import your own copy of each game; releases carry no game data.

    CrownParkComputingLanguage unknown★12

  • C nova-parcel A macOS menu bar widget for tracking Nova Poshta parcels in Ukraine, written in SwiftUI, single-purpose but honestly described, down to install steps. It only matters to Mac users of that courier, and the official app already covers most needs. #macOS#Productivity#APIs

    Unofficial macOS widget for Nova Poshta: parcel statuses, descriptions, and notifications. Swift / SwiftUI.

    FakerFoxJavaScript★7

  • C UniQuery4R A research repo claiming to unify 4D scene reconstruction in a single query, a novel topic within 4D reconstruction. But there's only a one-line abstract, with no method details, data or reproduction notes, so it's impossible to tell whether it's a real result or a placeholder. #ML Research#3D#Science

    UniQuery4R: Unified 4D Scene Reconstruction from a Single Query

    KosmoResearchPython★13

  • C gozarbin A Rust proxy client for OpenWrt routers aiming at lightweight, low overhead, but the one-line description omits supported protocols and how it differs from mihomo/sing-box, making it hard to judge whether switching is worthwhile. #Networking#Linux#Self-Hosted

    A rust core proxy client for openwrt routers

    dreamboxoneRust★8

  • C ps5-pkg-manager A package manager for PS5 aimed at homebrew users, a category that inevitably involves firmware versions and pkg distribution sources with murky legality; the repo also doesn't say who it's for or what prerequisites are needed. #Games#CLI#Hardware

    A clean and intuitive package manager for PlayStation 5.

    itsPLKC★8

  • C KnowYourConstellation Shows mission constellations and enemy forecasts early on the operations desk and briefing screens so players can gear up before deploying, a typical single-player experience mod. The scope is clear, but it depends on a specific game version and the author says nothing about compatibility. #Games#Plugins#UI Design

    Reveals mission constellations and enemy forecasts on the war table and briefing screen so you can choose your loadout before deployment.

    CowboyBingusLua★17

  • C voxel-beat-planet Turns any song into a Three.js voxel planet that bounces to the beat, visually appealing with readable code. But it's a one-off demo with no extensible interface, so reusing it in your own project would take major rework. #3D#Audio#Web Frontend#Animation

    Give any song a one-of-a-kind music-reactive Three.js voxel planet.

    EverettFishJavaScript★11

  • C nga-codex A userscript that disguises NGA forum pages as a Codex editor interface for slacking off at work. The need is real and it's fun, but it's pure front-end camouflage that falls apart up close, so its value stays at the joke level. #Browser Extensions#Social#Plugins

    A Tampermonkey script that disguises the nga forum as codex

    George11811JavaScript★8

  • C mobile-jev A mobile Jev project under the droidrun name with no description beyond the name, so it's impossible to tell whether it's a client, SDK or experimental code; not worth investing time for now. #AI Agents#LLMs#Android

    droidrunJavaScript★40

  • C blink Described in one line as 'Jev-powered codebase search' without explaining indexing, supported languages or integration cost, a thin wrapper. Repos branded with a model name have clustered lately, with no visible difference from existing semantic search; wait for a runnable demo. #AI#LLMs#Code Quality

    Codebase search powered by Jev from @typesafe-ai

    ellipsis-devTypeScript★11

  • C lyric-writing-skills A lyric-writing skill package for agents, essentially prompt and workflow assets whose effect depends entirely on the host model; similar skill collections have flooded in recently, with no verifiable quality claims. #Agent Skills#Prompting#Content Writing

    Professional agent skills for writing song lyrics

    jtydhr88Language unknown★18

  • C music-composition-skills A collection of composition and arrangement agent skills, again prompt assets, without saying which host tool they pair with or how output quality is measured; it's basically the same template as the previous batch of skill packages. #Agent Skills#Prompting#Audio

    Professional agent skills for composing and arranging popular music

    jtydhr88Language unknown★18

  • C CommandCodeGo-manager Wraps subscription accounts as OpenAI/Anthropic-compatible endpoints for reverse proxying, essentially bypassing subscription usage limits and concurrency constraints; it could be banned upstream at any time and is unsuitable for teams to depend on. #APIs#Networking#LLMs

    Command Code subscription reverse proxy: OpenAI/Anthropic-compatible endpoints + web admin interface + macOS/Windows desktop apps

    learningdog1JavaScript★10

  • C Nexus-Proxy A third-party proxy for Growtopia that calls itself "most advanced" but is actually mostly used for automation and cheating; the description is exaggerated and gives no protocol details, and several similar proxies already exist in the community without this one explaining any difference. #Games#Networking#Automation

    Most Advanced Growtopia Proxy

    markus6222C++★35

  • C effect-questions Packages semantic judgment as a single Effect primitive, an interesting direction, but the one-sentence description neither shows what the API looks like nor gives examples, so it cannot be installed or understood. It may be a teaser for Effect ecosystem users, but right now it is just a concept. #LLMs#SDKs#APIs#effect

    Semantic judgment as an Effect primitive.

    saiashirwadTypeScript★12

  • C jevpilot A Three.js driving simulator that uses Jev for autonomous driving decisions — an interesting visualization demo. But both the simulator and the model are only built to the point of running, so it's hard to take away anything reusable after watching. #Games#3D#LLMs#jev#three.js

    A playable Three.js driving simulator with Jev-powered autopilot

    standardagentsJavaScript★9

  • C ControllableHoverPack Adds "press space again to descend early" to a game's floating backpack while keeping the original landing assist, a restrained UX patch. The problem is it never names the game or gives install steps, so no one outside that game's players can judge its value. #Games#Plugins#UI Design

    Press Space again during hover-pack flight to descend early while preserving the pack's native landing assistance.

    CowboyBingusLua★8

  • C typesafe-mcp A thin wrapper packaging Jev as an MCP connector; MCP servers have proliferated in recent weeks and this offers nothing new in tool design, while the upstream model's credibility remains unverified. #MCP#AI Agents#LLMs

    mcp connector to give your AI agent direct access to typesafe ai's jev model

    itsmostafaGo★17

  • C anno117_pipe_example Apparently a pipeline sample for Anno 117 released by Ubisoft Mainz, useful reference for modders of that game. But the description is one sentence with no build or integration instructions, so for now it's just a browsable sample. #Game Dev#Games#Plugins

    Anno 117 - Pipe Example

    UbisoftMainzAnnoC++★7

  • C hermes-bot-forge A Hermes agent plugin that generates a complete bot with persona, memory, tools and gateway from one sentence. The idea is fun, but "complete, working Bot" is typical exaggeration, likely just template assembly, and it only works inside the Hermes ecosystem. #AI Agents#Agent Skills#Automation

    Hermes Agent plugin: say 'make me a <role>' and your agent builds a complete, working Bot — name, face, SOUL.md, memory, tools, skills, routines, gateway.

    BkashJEEPython★12

  • C pi-warden Adds guardrails to the Pi coding agent, using "Jev" to judge irreversible or off-track tool calls so people aren't interrupted. The direction makes sense, but the core judgment is fully outsourced to TypeSafe's scoring service, and the cluster of Jev repos in this batch makes it look like a joint promotion. #AI Agents#LLMs#Code Quality

    Guardrails for Pi built on pi-typesafe that steer the agent instead of interrupting you: Jev judges irreversible and off-task tool calls, detects stuck loops, checks unverified done claims, flags slop

    DevMortimerTypeScript★42

  • C prism2api A Python client reverse-engineering prism.openai.com's private API; it breaks whenever the protocol changes and carries terms-of-service risk. The description doesn't state usage boundaries, so beyond studying the API itself, no one should depend on it. #Reverse Engineering#APIs#LLMs

    Reverse-engineered Python API for prism.openai.com

    RobinfxaPython★20

  • C omarchy-hyprpwr-plugin Ports Hyper's HyperPower jitter effect to the Omarchy desktop, toggleable from the status bar. Simple to implement and honestly described, but it only serves users of a specific Hyprland distro, more fun than practical value. #Linux#Plugins#Animation

    Shake the whole Omarchy desktop like Hyper's HyperPower plugin. Toggle it from the bar.

    bjarneoQML★7

  • C qingzhu-sword-array An offline browser-playable pixel sword-flight fan game with original code under GPL, reasonably complete as a personal work; but there are no gameplay notes or maintenance commitment, limiting its audience to fellow fans. #Games#Web Frontend#Game Dev

    青竹剑阵 · 虚天殿篇: a pixel sword-flight fan game playable offline in the browser with no installation. Original code GPL-3.0-only.

    jindx1020-cryptoJavaScript★14

  • C awesome-jev A list of Jev-related projects and discussions that gained 23 stars in just over a day, showing the concept is gaining momentum. But the list is content aggregation with unfiltered entries; to understand Jev, reading the original implementations and evaluations is better. #LLMs#Knowledge Base#AI#jev

    A curated list of public projects, integrations, and discussions built on Jev — TypeSafe AI's System One model for typed decisions.

    yibiePython★23

  • C ai-landing-consultant Marketed as an open-source consulting system for AI agent deployment services, the repo has no reusable tools, only methodology packaging and delivery pitch. It may help clients seeking consulting, but for developers there's nothing runnable; it's more of a marketing page. #AI Agents#Automation#Marketing

    An open-source consultant system for deploying dedicated AI agents for specific users and landing their workflows | A consultant-in-a-box for deploying dedicated AI agents and shipping their first wor…

    August06exeShell★26

  • C farview-but-good The author self-deprecatingly calls it a silly packet-based view-distance extension, likely a Minecraft server mod. The honesty is refreshing, but a one-line description can't support a usability judgment, and beyond specific server owners it's unclear what it replaces. #Games#Plugins#Networking

    A silly packet based extended view plugin

    Taiyou06Java★7

  • C uplifting-biomolecular-modeling A biomolecular modeling repo under Anthropic's name with no description, so it's unclear whether it's paper code or leftover internal experiments; zero information beyond the name, not worth investing time based on it. #Biotech#Science#ML Research

    anthropicsPython★32

  • C building-with-jev-skill A skill package teaching agents how to write programs that call Jev, mostly prompts and specs; the repo is only hours old with nothing verifiable beyond concepts. #Agent Skills#LLMs#Prompting

    A skill for writing and improving programs that call Jev, TypeSafe's System One model

    dbreunigLanguage unknown★12

  • C DBlocker Claims to be a "second-layer ad blocker," but that one sentence is all there is — no telling whether it's DNS, proxy or browser level, and no install instructions. Ad blocking already has mature options like uBlock Origin; this doesn't even clarify its own positioning. #Privacy#Network Security#Browser Extensions

    An ads block layer 2

    thedusk4203JavaScript★17

  • C nautilo A self-hosted collaboration space billed as bringing AI into multiplayer, with a grand vision but no clarity on what it specifically does or where to start; typical big-platform narrative, unassessable at this stage. #AI Agents#Collaboration#Self-Hosted

    AI goes multiplayer. A self-hosted workspace for people and machine people. Create, code, and work together across desktop, mobile, and web. Open source. MIT licensed.

    agentseaTypeScript★23

  • C frida-ssl-bypass A Frida SSL unpinning script useful for mobile traffic capture, but several mature collections already exist and the repo has no description, so it's unclear what it adds over existing solutions. #Reverse Engineering#App Security#Network Security

    danieldev23JavaScript★18

  • C darrelltw-mods A self-described zero-token Claude Code 'mod' that also bundles Taiwan and US stock dashboards, a muddled positioning that is neither plugin nor prompt library, making actual usage hard to judge. #LLMs#Plugins#Productivity

    Claude Code mods by Darrell Wang — bands above the prompt, zero model tokens. Taiwan/US stock dashboard + more to come.

    darrell-twHTML★36

  • C anti-sycophancy-prompt A system prompt to curb model sycophancy, backed by 42 cited sources — more serious than similar prompt collections. But it's still text rather than a tool, its effect can't be verified, and readers can only take away wording inspiration. #Prompting#LLMs#AI

    A system prompt against AI sycophancy, backed by 42 sources on why models tend to agree with you.

    talhapythoneerLanguage unknown★7

  • C Google-Flow-Helper Rewrites API responses via remote rules to "unlock" Google Flow limits, which bypasses the provider's quotas, may stop working at any time, and carries account risk; not worth recommending to readers. #Crack#APIs#Video

    Unlocks Google Flow limitations by patching API responses based on remote specs.

    maanimeisamJavaScript★8

  • C POV-AI-Storytelling Billed as a practical framework for first-person AI visual stories, it reads more like a collection of prompts and methodology with no runnable components. AI short-drama makers might borrow a couple of workflow pages, but this isn't a tool and never explains how it differs from existing video workflows. #AI#Prompting#Video#Content Writing

    A practical framework for creating first-person AI visual stories.

    5263jhLanguage unknown★38

  • C PixelBoard Unlocks Gemini's voice dictation and writing features on any Android phone, essentially bypassing Google's feature and region restrictions. Tinkerers will be interested, but it relies on private APIs that can break anytime, and typos like "Google 11" in the description suggest it's unpolished. #Crack#Android#AI#Speech

    Unlock Google 11's Gemini "Rambler" Natural Voice Dictation & AI Writing Tools on ANY Android phone. No root required.

    AkshayykadamJava★32

  • C gta5-assets-helper Handles texture packing, metadata fixes and directory cleanup for GTA5 mods, real pain points for modders. But the repo is under a day old with a one-line description, no list of supported formats and no examples, so for now it's just a placeholder. #Games#Game Dev#Automation

    Manage, convert and validate GTA5 mod assets — texture packing, metadata fixes, and mod folder organization made easy.

    BarrieroacaptureLanguage unknown★12

  • C proxy-swarm-hub Claims to manage large-scale proxy pools with health checks and geo-rotation, but the repo is brand new with no implementation notes or usage examples, looking more like a placeholder than a runnable tool. #Networking#Automation#Self-Hosted

    Manage large proxy pools with health checks, rotation and geo-filtering — built for scraping and automation at scale.

    bondpigeonenergyLanguage unknown★12

  • C steepy-apex "Context engineering and governed workflows" for AI coding agents, all vision: project knowledge, expert routing, autonomous loops, but no concrete mechanisms, install steps or examples. Many agent-orchestration frameworks already exist, and it doesn't even state its differentiator. #AI Agents#LLMs#Automation

    Context engineering and governed workflows for AI coding agents. Keep project knowledge coherent, route work to specialists, and scale from small fixes to bounded autonomous loops.

    steepy-itJavaScript★8

  • C WorkBuddy-Switch2api An account switcher for WorkBuddy/CodeBuddy plus a local OpenAI-compatible reverse proxy, rotating an account pool and monitoring check-ins to farm free quota; it's also a fork, so compliance and stability are both concerns. #LLMs#APIs#Automation

    WorkBuddy / CodeBuddy account switching tool (Tauri + Rust) with a built-in OpenAI-compatible local reverse proxy: account pool rotation, credit check-in monitoring, and tray residency. Fork of change…

    cv-superdingRust★10

  • C mccccodereviewer An 'evidence-first' code reviewer that piles on terms like score-v2, regression radar, blast radius and What-If simulation, yet can't clearly say what goes in or out; heavy marketing tone, low credibility. #Code Quality#AI Agents#Testing

    MCC Code Reviewer — evidence-first repository intelligence with score-v2, regression radar, blast radius, codebase GPS, task lens, contract/test/architecture change analysis and safe What-If review si…

    darkseccPython★14

  • C musecrew An agent combo that produces multi-perspective research reports on meme coins on Robinhood Chain, essentially crypto-speculation content production; the claims of seven perspectives and traceable sources are backed by no concrete data sources. #Crypto#AI Agents#Blockchain

    musecrew — a plush research crew for Robinhood Chain memecoins. Seven perspectives, source-backed reports and a notebook that remembers.

    qwertyFragerTypeScript★13

DSkip101
  • D templatesgrokbot.com A library of 3000+ Grok bot templates accessed via MCP, solving the need to repeatedly write prompts when building AI assistants. It is essentially an SEO template site with low barriers, easily replaced once official template markets or prompt libraries mature, and its 51 stars mostly reflect traffic funneling rather than real usage. #Prompting#Marketing#MCP

    TemplatesGrokBot.com is a library of 3,000+ ready-made Grok Bot templates. Search by task or job title, install any bot in Grokbot in minutes, and connect the whole catalog to your agents via MCP. Mad…

    templatesgrokbotLanguage unknown★51

  • D unclutter This browser extension uses Jev to clean up cluttered web content by reusable template rules, aimed at tidying pages before reading or screenshots, which is not a new problem. Reading modes and ad blockers are already mature, its 46 stars are small, the model it relies on is untrustworthy, and it lacks a charging point. #Browser Extensions#AI Agents#Productivity

    WXT browser extension: Jev-powered page clutter removal with reusable template rules.

    kitzeTypeScript★46

  • D procyon An untethered re-jailbreak for 32-bit devices on iOS 10.0-10.3.4, showing real technical skill and solving jailbreak failures on old devices. But the target systems have been obsolete for years with a tiny audience and no commercialization space, making it only a technical reference for the jailbreak community. #Jailbreak#iOS#Reverse Engineering

    untethered re-jailbreak for iOS 10.0-10.3.4 (32bit)

    staturnzzC★50

  • D aniimo-modkit This is a 41-module Aniimo modifier offering combat survival, capture, and full-progress unlocks for cheating players who want to skip ahead; its 53 stars show the game does have this demand. But there is no legitimate commercialization path, it can be banned by official anti-cheat at any time, and it carries legal and account risks. #Game Cheats#Games

    41-module cheat/trainer for Aniimo — combat survival, Twining and elemental tools, creature capture, Homeland building and full progression unlocks.

    CoveBlackbirdFoldLanguage unknown★53

  • D buy-github-stars This is essentially a gray-market service manual for selling GitHub stars, addressing vanity metrics rather than engineering problems. Its 90 stars come from spectators and marketing traffic; the business exists but is platform abuse that GitHub may purge anytime and damages open-source credibility, so it should not be considered for investment. #Scams#Marketing

    Buy GitHub stars from real aged accounts - a complete guide to the delivery process, pacing, account quality and guarantees. Written for developers comparing star services before ordering.

    repoboost-hqLanguage unknown★90

  • D jev-codex-router A router that picks models and reasoning tiers for Codex turn by turn, essentially a promo piece for TypeSafe's Jev; several repos in the same batch carry the Jev name, so the star counts look unnatural. Interesting for multi-model users, but the value is tied to one paid service. #AI Agents#LLMs#CLI#SDKs

    Per-turn model & reasoning routing for Codex, driven by Jev (TypeSafe System One): picks the model, thinking depth and speed mode for every turn.

    0xNatoshiPython★8

  • D prismctl2api Yet another relay service wrapping GPT models behind its own API, described only as "access GPT via prismctl," with no sign of what it solves versus the many similar 2api projects. It depends on upstream accounts, and stability and compliance are unclear. #APIs#LLMs#DevOps

    Users just need to access the GPT model through prismctl.

    DanchunaGo★14

  • D jev-trader It makes an AI trading decision for every Monad block, solving no unique problem and looking more like riding an on-chain trend. Its 490 stars come from the same batch of repositories piling up TypeSafe Jev terms, with no sign of real users. Monetization is left to token issuance or meme plays, and the risk is that value drops to zero once the hype is exposed. #AI Agents#Crypto#Fintech

    One AI trade decision every Monad block. Jev on Kuru MON-USDC.

    jarrodwattsTypeScript★490

  • D ComfyUI-Luori-Sunset A ComfyUI plugin that acts as a prompt assistant for text-to-image, but one sentence does not clarify whether it does completion, translation, or templates; similar nodes are already numerous and it lacks differentiation. #Image Generation#Prompting#Plugins

    A prompt assistant for text-to-image.

    luori668Python★9

  • D fast-jev-compaction A Claude Code plugin that replaces default compression summaries with scoring, addressing the crude truncation of long-session context and targeting a real pain point. But it relies on the obscure Jev model, its 177 stars grew too fast, reliability and sustainability are questionable, and it lacks an independent monetization path. #AI Agents#LLMs#Plugins

    Claude Code plugin that replaces the compaction summary with Jev decisions: every tool call and result is scored in one fast request, stale ones are dropped or truncated, everything kept stays verbati…

    tamaratranTypeScript★177

  • D jevmeter Scores any video sentence by sentence with a "Jev score" and outputs a 16:9 cut, essentially a front-end demo for TypeSafe's scoring service; several repos in this batch share the same narrative, so credibility is discounted. Video-scoring folks can look at the approach, but it depends on that company's API. #Video#Video Editing#AI#LLMs

    Put a live Jev (TypeSafe) meter on any video: every sentence scored, rendered as a 16:9 edit

    ChetasLuaPython★32

  • D track-articulate-act Companion code for the paper Track, Articulate, Act, with no task definition, dataset or run instructions beyond the title, making reproduction hard for outside researchers and engineering value unclear. #Robotics#ML Research#Science

    Code for Track, Articulate, Act

    brains-bots-n-behaviorPython★7

  • D open-jev An experiment in typed JSON reasoning with DiffusionGemma, including benchmark results for "Every and Jev," essentially adding an evaluation piece to the Jev narrative. The method is worth a look, but it belongs to this batch of clustered related repos, so its independence is doubtful. #LLMs#ML Research#Benchmarks#AI

    Typed JSON inference with DiffusionGemma, with Every and Jev benchmark results

    JoshuaSPPython★9

  • D funnel-mkt A Vietnamese marketing-funnel scheme using 8 Claude Code agents split by a "pizza model" to pull leads — essentially prompts plus a process template. It may save effort for local ad teams, but for open-source readers it's neither reusable nor easy to verify. #AI Agents#Marketing#Prompting#claude-code

    He thong 8 AI agent chay theo mo hinh banh da de xay pheu marketing thu lead. Viet cho Claude Code, tieng Viet, VND.

    thaofvn-coca06HTML★11

  • D foreman Claims to be a software factory scheduler built on TypeSafe Jev, but the description is vague and never specifies what it schedules or how outputs are verified. Its 187 stars appeared the same day as a batch of other Jev repositories with no evidence of real use, and the commercialization path is unclear, likely making it a bandwagon concept repo. #AI Agents#Agent Skills

    Software factory foreman based on TypeSafe's Jev model

    thruwirePython★187

  • D jev-eval-agent 59 stars but not a single line of description, so it's impossible to tell what it evaluates, how to run it, or where results are. Jev-related repos have appeared densely within two days with similar star counts; this empty shell looks like a hype placeholder — wait for documentation. #LLMs#Benchmarks#AI#jev

    vinilanaHTML★59

  • D awesome-cissp CISSP cram notes for security architects, leaning toward business-side risk, governance and legal perspectives, with a clear audience. But it's a pure content repo whose value is the author's writing rather than code, so it doesn't belong on an open-source project list. #Learning#App Security#Knowledge Base

    A last-mile CISSP guide for security architects: the business side of cyber (risk, governance, law, and the manager mindset)

    vitalsecurityLanguage unknown★19

  • D DeadOps A GSC script loader for jailbroken PS4/5 with a built-in BO2 GSC compiler, whose real use is basically script modding. It has a high technical bar and a tiny audience, and adding scripts on consoles risks bans and legal boundaries. #Game Cheats#Reverse Engineering#Games#Exploits

    A GSC loader SPRX for PS4/5 with a built-in BO2 GSC compiler.

    ItsJokerZzC++★12

  • D tg-stars-autofill An automation script for bulk-buying Telegram Stars and managing balances, with no clarity on funding sources or API compliance and very thin docs. Bulk top-up tools carry risks far outweighing convenience, so there's no reason to spend time. #Automation#Social#Fintech

    Automate Telegram Stars top-ups and balance management — batch processing, retry logic, and clean reporting.

    SeparateiashedLanguage unknown★12

  • D CS2-DMA A CS2 DMA memory-reading framework framed as research, covering entity state parsing, PCIe experiments and synchronization, but it is still cheat infrastructure and the academic packaging doesn't hide its purpose. #Game Cheats#Reverse Engineering#Windows

    Open-source, low-level Direct Memory Access (DMA) research framework for Counter-Strike 2, focused on memory acquisition, entity-state parsing, synchronization, PCIe experimentation, and reverse-engin…

    cs2-dmaC++★17

  • D viperlith Packages Datastar's hypermedia ideas into a Python monolith framework for backends wanting less JS, but the README has only emoji and a slogan, no install, examples or scope. Versus FastAPI plus Datastar by hand, no increment is visible; wait for a runnable example. #Web Frontend#SDKs#datastar

    🐍⚡🚀 Viperlith: the Hypermedia-Based Monolith for Python built with Datastar

    eliasdejongPython★13

  • D Seed-Phrase-Generator A Bitcoin and Ethereum mnemonic brute-force and collision tool claiming to recover wallets, but in substance a phishing-style gray business exploiting greed. Collision hit probability is negligible, its 51 stars are speculation-driven, it has no legitimate commercial value, and it carries clear legal and fraud risks. #Crypto#Crack#Scams

    WalletGen - Seed Phrase generator / checker for Bitcoin (BTC) and ETH (Ethereum) wallets. Find or recovery your wallet via brute force and brain wallet generation

    tony-devbtcC++★51

  • D iWork A four-word autonomous office agent with no capability boundaries, input/output examples or run instructions, a repo with slogans but no usage. #AI Agents#Productivity#Automation

    Autonomous office agent

    ai-4WorkPython★8

  • D jev-review This is a staged code review workflow and local panel built on TypeSafe Jev, aiming for an AI review loop, but it depends on a model with no credibility, so its value cannot stand on its own. Its 160 stars appeared in sync with the same batch of repositories, suggesting inflated numbers, and the commercialization outlook is unclear. #AI Agents#Code Quality#MCP

    A staged code-review workflow and local dashboard built with TypeSafe Jev.

    devagrawal09TypeScript★160

  • D typesafe-ai-benchmark A joke LLM gateway repo calling itself a 'fake Jev', with no visible difference from structured-output gateways like LiteLLM and no usable configuration docs, so it's only good for a laugh. #LLMs#APIs#Benchmarks

    This is a LLM Gateway that mimics typesafe ai structured output. Like an imposter Jev.

    iammrduncanTypeScript★25

  • D moegpi The author packaged their Pi coding agent skills, Exa search and model preferences into an installer, open-sourcing a personal setup. Like-minded users can copy the environment, but it has no general value and doesn't solve any new problem. #AI Agents#CLI#Productivity

    My Pi coding agent setup: skills, Exa search, model preferences, and an interactive installer.

    MOEG-5Python★8

  • D typesafe-mario It lets an agent play Super Mario using structured simulator state, a demo-style project that solves no real engineering problem. Its 205 stars surged in sync with the same batch of TypeSafe Jev repositories, looking more like coordinated hype than organic growth. There is no commercialization path, and it will likely go unmaintained once the hype fades. #AI Agents#Games#LLMs

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

    fhshaikPython★205

  • D friendlilink Says only 'a convenience tool for Friendli integration', without stating what services it integrates, who it's for, or any usage or examples. External developers can't judge whether to install it. #SDKs#APIs#AI

    For FriendliLink, our integration convenience tool.

    friendliaiTypeScript★8

  • D jev-router It routes requests within Claude Code to the cheapest model per task to save money, a practical idea, but many LLM routing solutions already exist. With 64 stars and reliance on a Jev model of unknown origin, the implementation is hard to verify and there is no sustainable charging point. #AI Agents#LLMs

    Route to the cheapest model in claude code for your task using jev-router

    gargpratyushJavaScript★64

  • D jev-review This is a local-first MCP plugin using Jev for continuous code quality review. Its functionality heavily overlaps with the similarly named jev-review, showing this wave is the same concept cloned in bulk. Its 75 stars lack evidence of real users, the core model it depends on has no credibility, and it lacks independent commercial value. #AI Agents#Code Quality#MCP

    Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.

    NiazMorshed2007TypeScript★75

  • D qwen38-flash-next-w4a16-cmp170hx The model name Qwen3.8-Flash-Next in the title does not exist, yet the author piles on tok/s and concurrency numbers, looking like a batch-generated degenerate benchmark repo whose figures cannot be verified. #LLMs#GPU#Performance

    Qwen3.8-Flash-Next (W4A16) on 2x CMP 170HX: 110 tok/s decode, 7K prefill, 800 tok/s @ 23 concurrent - runs on 38 GB home RAM, no pinned memory

    polyuij42-delPython★9

  • D lighter-arb An undescribed Python arbitrage script with no indication of which exchange it targets and no risk controls, backtesting or fund-safety notes. Copying and running such code carries risks far exceeding the returns. #Crypto#Fintech#Automation

    UngusCodePython★7

  • D china-historical-visual-resources An index of Chinese historical imagery, a pure content collection with no tools or code output; convenient perhaps for researchers, but not an open-source project worth reviewing. #Learning#Documents#Knowledge Base

    An index of Chinese historical image materials: selected photography, collections, and special topics, with Chinese guides, categories, and sample materials.

    boryzooLanguage unknown★8

  • D jev-browser A Jev-driven element selector for existing agent computer-use tools, a small link in the computer-use pipeline that lacks independent value. Its 44 stars are part of the same Jev hype cluster with no evidence of real use or commercialization path, so skip it. #AI Agents#Web Frontend

    Jev-powered element selection for your agent’s existing computer-use tools

    vlad-terinJavaScript★44

  • D awesome-typesafe This is a resource list around TypeSafe, System One, and Jev, essentially self-justifying material for this wave of coordinated hype: first coin terms, then use a list to endorse them. It has no technical substance or monetization of its own and serves only as a negative signal about the credibility of this cluster. #AI Agents#Knowledge Base

    A curated list of official resources and community projects for TypeSafe, System One models, and Jev.

    AbdelStarkCSS★65

  • D Among-Us-Tool An Among Us enhancement or cheat tool, written like a typical hack repo: all adjectives, no implementation, and bound to be blocked by the official side; no reader need touch it. #Game Cheats#Games#Windows

    Advanced and stable enhancement tool for Among Us. Designed for better gameplay experience with clean interface, useful features, high performance and regular updates. Easy to install and use on the l…

    TabooMarshalUncoverLanguage unknown★27

  • D video-analyzer No description; Python video analysis could range from frame sampling statistics to multimodal understanding, and the repo sets no boundaries, so it can't be evaluated. #Video#AI#Multimodal

    knishika62Python★11

  • D sing-box-config-template A sing-box config template, but the repo has no description and doesn't say what scenario it targets or how to apply it. Similar samples are readily available in the community; this one offers no trade-off explanation or validation, so it's impossible to judge whether it's easier. #Networking#Privacy#Self-Hosted

    wuziweiLanguage unknown★10

  • D ios.CVE-2026-84616-84607 It claims to trace the causes and fixes of two iOS CVEs, but the numbering looks dubious and the content is nearly empty, so correspondence to real vulnerabilities can't be verified. Vulnerability research is worth reading only with reproducible details, which are absent here. #Exploits#iOS#App Security

    research on finding the bug and fix of CVE-2026-84616 and CVE-2026-84607

    Ping-2oLanguage unknown★10

  • D Xing4.0-29B-A4B A model repo containing only a model name, with no training data, evaluation, license or inference instructions, and a name like 29B-A4B that can't be verified. It looks more like a placeholder or name squat than a usable release. #LLMs#ML Research#Benchmarks

    XingChen-AGILanguage unknown★27

  • D jev-mcp This is an MCP proof of concept for the so-called TypeSafe Jev model, solving no concrete problem beyond wiring a new term into MCP. Its 46 stars appear among the same batch of coordinated-hype repositories, with no real users or signs of sustainable maintenance, so it has no investment value. #MCP#AI Agents

    Fast, cheap, typed judgments from TypeSafe's Jev model, as MCP tools.

    jkudishTypeScript★46

  • D family-school-cards No description; the name suggests a homeschool card app, but the repo has no usage notes or screenshots, so it's unclear whether it's finished or a draft. #Learning#Documents#Web Frontend

    jot-koehlerJavaScript★9

  • D smartfactoryAi No description; named smart factory, but it's unclear whether it's a simulation, dashboard or course demo, and HTML pages make completeness hard to judge. Insufficient information. #IoT#Data Analysis#Web Frontend

    kdn986532HTML★11

  • D nvidia-register Uses Playwright to mass-register NVIDIA BUILD accounts, nominally for "learning automated testing" but actually to farm free credits, violating the terms of service; not recommended for anyone. #Scams#Automation#AI

    Automated NVIDIA BUILD account registration toolkit (Playwright) — for personal learning of automation testing

    qiukui666Python★8

  • D jev-drone It uses a small Jev decision model to closed-loop control a vision-only drone in MuJoCo, a demo-style research project whose 53 stars come from the same wave of Jev concept inflation. It has neither real robot validation nor a commercialization scenario, giving it only demonstrative value. #Robotics#AI Agents#ML Research

    Camera-only autonomous drone in MuJoCo with a small judgment model (TypeSafe Jev) in the loop at 2.5Hz

    RomanSlackPython★53

  • D proxy-config-templates No description, and the repo name alone doesn't reveal which client it targets or whether nodes are included, likely just a personal config backup. Insufficient information to judge value. #Networking#Self-Hosted#Privacy

    forechoLanguage unknown★40

  • D harness No description, a Shell repo with overly generic filenames, so it is impossible to tell whether it is a test fixture or deployment scripts. Skipped per the rules. #Automation#DevOps#CI

    natevickShell★13

  • D qwen38-flashnext-exl3 No description, and the repo name also carries a nonexistent Qwen model, suspected to be from the same batch of templated deployment repos; impossible to evaluate. #LLMs#GPU#ML Research

    r0b0tlabPython★16

  • D mario-jev The repo has no description, and from the name it is a small demo of Jev playing Mario. It is impossible to tell what it does or who it is for, so it is out per the rules; demos of models playing classic games are already plentiful, and missing this one changes nothing. #Games#LLMs#AI#jev

    shantanugoelPython★7

  • D RimAssistantLimbusCompany The repo has no description; the name suggests a RimWorld assistant plugin, but function and usage can't be judged. With 16 stars within a day of creation and no docs, it looks hastily made; wait for a README. #Games#Plugins#Automation

    RatioRatPierLanguage unknown★16

  • D super-access-analytics-public An undescribed Python repo whose name hints at access analytics, but data sources, deployment and output format are all invisible; it looks like a public slice of internal code with nothing to evaluate. #Data Analysis#APIs#Monitoring

    TigerMondayPython★8

  • D yt-marketing-bot A YouTube view-boosting framework that auto-subscribes, likes and simulates watching on sub4sub sites, which is platform cheating; the docs describe only the flow, not the risks, and no one should use it. #Scams#Social#Automation

    A Python framework autonomously handles YouTube channel subscriptions video likes and simulated views on sub4sub websites.

    connorcruz-24Language unknown★10

  • D awesome-jev An awesome list of links around the so-called Jev model, whose source model is itself unverified, and the list has no selection criteria. A trend-chasing directory repo with no reading value. #AI#LLMs#Knowledge Base

    A community directory of projects built on Jev, TypeSafe AI's System One model.

    hellogumboJavaScript★8

  • D TCDD No description, not even what the repo name abbreviates, and the Python code could range from research scripts to small tools. Impossible to tell what problem it solves. #AI#ML Research#Learning

    hunkimPython★8

  • D projet-cdc-neurolink No description, only a project name, so it is impossible to tell whether this is coursework, a graduation project, or a real product. Skipped per the rules. #Health#Biotech#Science

    le-reboursLanguage unknown★10

  • D github-search-ranking Claims it can push a repo into GitHub search top five for any keyword, a black-hat growth service unrelated to open source itself, with no method explanation and only a service for sale. #Marketing#Scams

    Rank your repository in GitHub search - top 1-5 on any keyword you name, globally, with demos of achieved ranks on request.

    repoboost-hqLanguage unknown★36

  • D Vehicle-Sales-Visualization-Platform The name suggests car sales visualization, the description is empty, and it looks like coursework or personal practice, with no reusable components or engineering decisions worth learning from. #Data Analysis#Web Frontend#Learning

    Yawn62Python★13

  • D vulnerability-writeups Named as a vulnerability analysis collection, but the repo has no description and no content with discernible structure, so it's impossible to tell original research from a repost digest. An empty shell offers no reason to recommend it; wait until it at least lists its articles. #App Security#Exploits#Learning

    taohuams-pinkLanguage unknown★9

  • D APple-gAmE The description itself says it's "just a ChatGPT mini-game, good enough," with no hint of what you play or how, and no readable docs. It's a personal practice project with loose syntax and naming, not worth readers' time. #AI#Games#Web Frontend

    It's just simply chat gpt ai game It's good enough

    Cheon-8642HTML★11

  • D Snake An undescribed HTML Snake game, a standard practice assignment with neither gameplay innovation nor engineering showcase value. #Games#Web Frontend#Learning

    MalicLOalHTML★9

  • D home-assistant-lidlplus No README; the name suggests a Lidl Plus Home Assistant integration, but neither purpose nor configuration is documented. It might be valuable to HA users, but there's no way to judge yet, so wait for the author to add notes. #Self-Hosted#APIs#Automation

    Przemko92Python★9

  • D Fortnite-Cheat-2026-External-Aimbot-ESP A Fortnite cheat; the for educational purposes label is standard disclaimer language, while it actually provides aimbot, ESP and anti-detection toggles, a repo that clearly should not be recommended. #Game Cheats#Games#Windows

    Fortnite cheat 2026 with aimbot, ESP, wallhack and configurable settings for Windows 10/11. Undetected external tool for educational purposes.

    calmserferadicateLanguage unknown★11

  • D goauntlet An undescribed, undocumented JavaScript repo whose name hints at a game or challenge project, but there are no clues to what it does or who it's for. #Games#Web Frontend#Learning

    codetesla51JavaScript★8

  • D rivals-mod-menu A Rivals mod menu with only a title, no feature description or implementation details, making it the most common filler in cheat repositories. #Game Cheats#Games#Windows

    Ultimate Rivals Menu

    cookc8441Language unknown★18

  • D cs-2-silent-menu A CS2 silent aim menu marketed as automatic bullet snapping with a fixed crosshair, a typical cheat tool with no technical value beyond evading detection. #Game Cheats#Games#Windows

    CS2 silent MOD MENU — bullets snap to targets without visible crosshair movement.

    coxt-8150Language unknown★10

  • D GOT-kingsroad-mod A mod menu for Game of Thrones: Kingsroad with nothing beyond a title and year, a typical cheat repository worth skipping. #Game Cheats#Games#Windows

    Game Of Thrones Kingsroad Mod Menu

    danielcole90Language unknown★10

  • D jev-me The description is just one line, 'chat with Jev', with no clear features or inputs and outputs, not even whether it's a tool or a script. Can't judge its purpose. #AI#LLMs#AI Agents

    Grill-me with Jev optional each turn

    jon-devlapazLanguage unknown★9

  • D learning-python A personal Python learning repository that is neither a tool nor solves any new problem, with 98 stars coming mostly from social attention rather than practical value. It has no potential for commercialization or becoming a dependency, so skip it and don't be misled by the star count. #Learning

    My Python learning journey — basic programs, practice, and projects.

    sabbircode-devPython★98

  • D xingcezhenti A packaged repo of Chinese civil service exam questions from 2016–2026, pure content collection that also involves copyright issues. It has reference value for test takers, but it isn't an open-source project and doesn't belong on a tech list. #Learning#Documents#Knowledge Base

    All 行测 past exam questions from China's 2016-2026 national and provincial civil service exams

    ERRRCLanguage unknown★14

  • D Valheim-Ultimate-Offline-Trainer-Collection A collection of offline cheats for Valheim, with god mode, infinite stamina and flight, plus a claim of "zero ban risk." This is a pack of cheat resources with no original technology, and it shouldn't be recommended as an open-source project. #Game Cheats#Games#Plugins

    🛡️ Ultimate offline trainer & mod collection for Valheim 1.0. God Mode, Infinite Stamina, Unlimited Items, Fly Mode, No Crafting Requirements, Map Reveal, and more. Features FLiNG, Fearless Revolution…

    Janne1157Language unknown★19

  • D metin-2-mod A Metin 2 menu unlocker for players wanting to modify the client; the repo explains no principle, supported versions or ban risk, just an unlocker name, a typical cheat utility not worth ten minutes given the many similar injection scripts on forums. #Game Cheats#Games#metin2

    Metin 2 Mod mEnu unlocker

    davidjimenez60Language unknown★11

  • D sij An undescribed, undocumented HTML repo with no visible problem it solves or audience, and 9 stars in 1.6 days offer no verifiable clues; not worth opening now. #Web Frontend#Documents#Productivity

    SeongImJuHTML★9

  • D project-zomboid-menu A Project Zomboid mod menu with only a title and phrases like '2026 latest', no implementation notes, typical filler in cheat listings. #Game Cheats#Games#Windows

    PROJECT ZOMBOID MENU MOD 2026 NEWEST

    danielm-7134Language unknown★10

  • D HCL-Java-Full-Stack-Training A collection of class notes and assignments from a Java full-stack training course, meaningful only to fellow students, with no reusable tools or methods. It got 12 stars one day after creation, with neither original content nor documentation structure, a typical content repo not worth a leaderboard slot. #Learning#Knowledge Base#Documents

    Training notes, study materials, Java Full Stack programs, assignments, projects and resources for HCL Java Full Stack Training.

    rgwonder786Java★12

  • D mc-account-recovery Claims to recover and migrate Minecraft accounts and transfer skins, but such "account recovery" tools are usually either ineffective or bait to steal credentials. With no implementation details, it offers players only risk and no value. #Scams#Games#Privacy

    Recover and migrate Minecraft accounts — restore access, transfer skins, and back up profiles safely.

    DaimyoSheikhLanguage unknown★12

  • D minigame The description says only 'simple game', hand-written HTML, with no gameplay notes, screenshots or run instructions, a personal practice project offering nothing other developers can judge or reuse. #Games#Game Dev#Web Frontend

    simple game

    eric8966HTML★12

  • D 2d-vtuber-vtube-studio-setup-toolkit Claims to be an all-in-one configuration pack covering Live2D models, face tracking, hotkeys, and OBS scenes, but is actually a code-free compilation of instructions with an enormous scope and nothing explained in depth; the official VTube Studio docs already cover most of it, with no added value here. #Streaming#Animation#3D#vtube-studio

    2D VTuber & VTube Studio setup toolkit for Live2D avatars, face tracking, expressions, hotkeys, OBS scenes and streaming configuration.

    samashaw5861Language unknown★9

  • D lotto-generator A single-page tool that generates Korean Lotto numbers (with bonus); the logic is just random numbers, with no prediction and no data API. It's a personal toy any language can replace in ten lines, not an open-source project worth attention. #Games#Automation#Productivity

    Generates even the lotto bonus number

    strsrtHTML★11

  • D AGENTS.md The author's personal AGENTS.md config, a private workflow file rather than a reusable tool. Copying a couple of prompt lines might help someone, but as a project it has no engineering content and never explains why it's written this way. #AI Agents#Prompting#Agent Skills

    my AGENTS.md

    Da1sypetalsLanguage unknown★29

  • D bodycam-hack-2026-combat-toolkit Named as a Bodycam helper tool, it is actually another copy of a batch cheat template: the feature list, wording, and fields are nearly identical to same-named repos from the same period, with no runnable code or technical explanation, zero value to developers, and only likely to induce downloads. #Game Cheats#Crack#Scams#Games#bodycam

    Bodycam Hack 2026 themed PC companion for weapon loadouts, graphics and sensitivity profiles, maps, aim practice, match tracking and combat analytics.

    russellgordon-65Language unknown★9

  • D civilization-vii-cheat-strategy-toolkit A so-called strategy companion riding the Civilization VII name, whose README is only a checklist of leaders, techs, economy, and so on, with neither implementation nor data sources, from the same batch of game-hack templates seen before, existing purely for search traffic. #Game Cheats#Crack#Scams#Games#civilization-vii

    Sid Meier's Civilization VII Cheat themed companion for leaders, civilizations, cities, technology planning, economy, diplomacy and strategy tracking.

    ryanm-4870Language unknown★10

  • D star-wars-zero-company-hack-tactical-toolkit A tactical companion riding the Star Wars: Zero Company name, whose content is only feature entries for squads, equipment, and campaign progress, with no implementation or data sources, the same template as the batch of hack series from the same period, star-farming for traffic. #Game Cheats#Crack#Scams#Games

    Star Wars Zero Company Hack themed PC companion for squad building, character profiles, tactical missions, equipment and campaign progression.

    sandrasanders322Language unknown★10

  • D overwatch-2-hack-2026-hero-combat-toolkit An Overwatch 2 "hero data + aim practice" page, a traffic-farming repo mass-produced from the same template: no code, no data, no sources, with hack keywords slipped into the description. It offers no information value to developers. #Game Cheats#Crack#Scams#overwatch

    Overwatch 2 Hack 2026 themed PC companion for hero profiles, team compositions, aim practice, maps, settings and competitive match analytics.

    sherrera977Language unknown★10

  • D fivem-hack-2026-gta-v-roleplay-toolkit A supposed FiveM roleplay server helper that only contains server data, vehicle lists and "analysis" fields with nothing runnable — a template repo riding FiveM hype, from the same batch as the mass-produced cheat series. #Game Cheats#Crack#Scams#Games#fivem

    FiveM Hack 2026 themed GTA V companion for server profiles, vehicles, character builds, maps, roleplay planning and session analytics.

    sierrawood01Language unknown★10

  • D cs2-hack-2026-premier-combat-toolkit A CS2 "weapon profiles + aim practice" page, a templated product: a long feature list, zero implementation, not even data sources. The "hack" in the name only serves search traffic and isn't worth opening for any developer. #Game Cheats#Crack#Scams#cs2

    CS2 Hack 2026 themed PC companion for weapon profiles, grenade notes, maps, economy planning, manual aim practice and Premier match analytics.

    spencerc-5067Language unknown★9

  • D marvel-rivals-hack-2026-ranked-hero-toolkit A Marvel Rivals "hero profiles + team comps + aim practice" page, word-for-word identical to a dozen cheat templates from the same period: no code, no sources, a feature list built only to stuff keywords. Developers opening it will just waste time. #Game Cheats#Crack#Scams#marvel-rivals

    Marvel Rivals Hack 2026 themed PC companion for hero profiles, team compositions, abilities, manual aim practice, maps and ranked analytics

    timocox2996Language unknown★9

  • D pubg-hack-2026-ranked-survival-toolkit A PUBG drop spots, loot and "aim practice" list — another copy of the templated cheat page, with no implementation or data source. The 2026 and hack in the name exist only for search ranking and offer zero value to developers. #Game Cheats#Crack#Scams#pubg

    PUBG Hack 2026 themed PC companion for weapon loadouts, drop planning, rotations, inventory profiles, manual aim practice and ranked analytics.

    tlee6354Language unknown★10

  • D destiny-2-hack-2026-guardian-build-toolkit A Destiny 2 guardian build and activity list, same structure as the mass cheat templates: no code, no data sources, feature entries only for keyword stuffing. For builds, the community wiki is better; there's nothing usable here. #Game Cheats#Crack#Scams#destiny-2

    Destiny 2 Hack 2026 themed PC companion for Guardian builds, weapons, subclasses, activities, Crucible practice and progression analytics

    valeriam-7519Language unknown★9

  • D tf2-hack-2026-class-combat-toolkit A Team Fortress 2 class and weapon list, another hack funnel page from the same template batch, with zero implementation and zero sources. These repos appeared en masse in the past two days with similar star counts, suggesting star farming rather than real interest. #Game Cheats#Crack#Scams#tf2

    Team Fortress 2 Hack 2026 themed PC companion for class profiles, weapon loadouts, maps, manual aim practice and match analytics.

    vict-gibson02Language unknown★10

  • D deadlock-hack-2026-hero-build-toolkit A Deadlock build and lane list, still the same batch of cheat templates: only itemized descriptions, nothing runnable, no data sources. It carries no information for developers and has nothing to do with real tools. #Game Cheats#Crack#Scams#deadlock

    Deadlock Hack 2026 themed PC companion for hero builds, item planning, lanes, manual aim practice, team compositions and match analytics

    victor-2933Language unknown★9

  • D battlefield-2042-hack-2026-specialist-combat-toolkit A Battlefield 2042 specialist and vehicle list, yet another templated cheat page: no implementation, no data sources, feature lists only for keyword coverage. Like the dozen isomorphic repos from the same period, it isn't worth opening. #Game Cheats#Crack#Scams#battlefield-2042

    Battlefield 2042 Hack 2026 themed PC companion for specialist profiles, weapon loadouts, vehicles, maps, manual aim practice and multiplayer analytics.

    walterb-8961Language unknown★8

  • D xiaoxiao-jacoco The repo has no description; from the name it looks like a personal experiment related to JaCoCo coverage, and it doesn't even mark the language. There's no telling what it does or who it's for today, and no entry point to run it, so we can only wait for the author to add a README. #Testing#Code Quality#Build

    000xiaoxiao000Language unknown★13

  • D inputcn No README and no description, so it's unclear whether this is an input method, an input component, or something else. The cn in the name is the only clue, which gives readers nothing to evaluate. #Web Frontend#UI Design#Productivity

    42aditya31TypeScript★9

  • D optibuild-showcase No description, and the showcase in the name suggests it's just a product showcase page; the repo itself contains nothing to evaluate. Readers can't tell what it does or where to start. #DevOps#Build#Marketing

    AIEngineer0923Language unknown★9

  • D temasuite-showcase Like the author's other showcase repo, there's no explanation at all and it looks like an empty shell for product promotion. Nothing runnable, and no indication of which existing tools it corresponds to. #Marketing#UI Design#Web Frontend

    AIEngineer0923Language unknown★7

  • D Agent_Ai No description or docs at all; from the name one can only guess it's AI agent related, with no way to tell what it does or who it's for. There's nothing in the directory to evaluate. #AI Agents#AI#Web Frontend

    Crowtit517HTML★9

  • D moonlight_calendar No README and no description, so it's unclear whether this is a calendar app or something else; it appeared the same day as another empty repo by the same author, looking more like a quick placeholder. Readers have no way to judge its purpose. #Web Frontend#Productivity#UI Design

    Crowtit517HTML★9

  • D SkillGate No description at all; the name hints at agent skill admission, but there's no explanation or usage in the repo. Readers can't tell what problem it solves. #AI Agents#Agent Skills#Code Quality

    Lihao-LeoJava★8

  • D Aula_AiGen The description contains only a name and email, with no feature explanation or code clues in the repo, so its purpose can't be determined; treat it as undocumented. #AI#Prompting#Content Writing

    Campari Mirko - [email protected]

    MaSTERmIKKLanguage unknown★7

  • D aigccat No description or explanation; the name only hints at something AI content related, so what it does, who it's for, and any assessable implementation remain unknown. #AI#Content Writing#Automation

    RainNamelessPython★8

  • D Multichain-Drainer The description plainly states it's a cross-chain wallet drainer for emptying ETH, USDT and other assets, padded with keyword spam. Such a repo shouldn't be recommended regardless of completeness; it's only useful as a threat-intelligence sample. #Scams#Crypto#Blockchain

    A bot or script designed to steal assets from cryptocurrency wallets across multiple blockchains—such as Ethereum (ETH), Tether (USDT), and various other tokens or NFTs—is commonly known in the cybers…

    PegasusivitalizeLanguage unknown★11

Show HN 5

A

Show HN: MCPJam - the first testing & evaluations platform for MCP servers

Show HN

A testing and evaluation platform for MCP servers, addressing the inability to observe MCP services in agent clients or run regression tests; the founder comes from Asana and brings an open-source user base as momentum. The commercialization path is clear: per-seat CI integration pricing, with the risk of being squeezed by Anthropic's official Inspector and various free evaluation tools.

Prathmesh, CEO of MCPJam here. Users now start in ChatGPT, Claude, Cursor, and other AI clients. They reach your product through your MCP server. That means your users often aren’t in your product any…

#MCP#Testing#AI Agents#Code Quality

▲9

A

Show HN: I built a new version of my fun spatial 3D online meeting app

Show HN

Turns remote meetings into a multiplayer game where you walk around a 3D space, addressing the tedium and lack of serendipity in video calls; rebuilt solo over years, its 92 points indicate real users. Commercialization means team subscriptions, with risk from competitors like Gather and the ongoing operating costs of 3D scenes.

Hi HN! flat.social is a fun spatial online meeting app for remote teams and communities. It's largely built like a web multiplayer game, and the 3D virtual spaces are highly customisable with elements…

#Collaboration#3D#Web Frontend

▲92

B

Show HN: Aclif – Agent CLI framework: one grammar, canonical names across SaaS

Show HN

Unifies various SaaS CLIs under one syntax and naming convention so agents can call different services with the same commands, addressing fragmented tool invocation; its 29 points were relatively high that day, and default schemas can be discovered without credentials. It could become a standard interface for the agent tool layer, with the risk that vendors may not maintain the adapters.

Simple discovery of the provider's default schemas is possible without any credentials (since they are built into the binary). If you want it to fetch customizations of your specific instance, give it…

#CLI#AI Agents#APIs

▲29

B

Show HN: OJ – A drop-in replacement for Vite in Rust

Show HN

Rewrites Vite natively in Rust to fix slow React builds and cold starts; 399 stars in one day shows real developer frustration with Vite's performance limits. Commercialization means acquisition by a big company or becoming React's default build chain, but the biggest risk is that VoidZero's Rolldown/Rspack already hold the ecosystem slot and plugin compatibility is very hard to catch up on.

#Build#Web Frontend#Performance

▲7

B

Show HN: Craigslist for agent skills, curated by a human

Show HN

Turns agent skills into a marketplace where ordinary people can buy ready-made capabilities instead of writing them; its 20 points show HN curiosity about skill trading. Commercialization means a commission marketplace, but cold-starting both supply and demand is extremely hard, and the strongest competitors are Anthropic's official skill catalog and free GitHub collections.

hey, i'm nick (@skeptrune on X). i have been using ai for a lot of things i'm not an expert in and thought it would be nice to try and buy paid skills for helping it accomplish those tasks more effici…

#AI Agents#Agent Skills#E-commerce

▲20

Hot New Repos

  1. browser-use/jev-ultrafast★1984
  2. TheoLeeCJ/openjev★749 Can we run something like Jev on a 3090 at home?
  3. vinnylarouge/jevlike★654
  4. jarrodwatts/jev-trader★490 One AI trade decision every Monad block. Jev on Kuru MON-USDC.
  5. ctdal/cve-2026-41940-PoC★378 A cPanel and WHM authentication bypassing tool
  6. korcarc/text-humanizer★365 text-humanizer is an open-source project designed to convert text generated by AI to its "humanized" version. Bypasses the most of AI detectors such as Turnitin or GPTZero
  7. Marcos66236/github-stars-history★276 Track and visualize the stars history of any GitHub repository. Open-source growth analytics and velocity tracking.
  8. Mak5er/AirCard★270 Apple Wallet Card Skinner for iOS 18+ (No Jailbreak Required)
  9. fhshaik/typesafe-mario★205 A TypeSafe/Jev agent that plays Super Mario Bros. from structured emulator state.
  10. yynxxxxx/gpt_sub_analysis★199 Security analysis of the ChatGPT iOS subscription / entitlement flow (internal draft, coordinated disclosure WIP)

Star gain 1d

  1. cloudflare/security-audit-skill+2990★8653
  2. Tencent/WeKnora+931★25649
  3. TheoLeeCJ/openjev+700★749
  4. anthropics/claude-code+588★145261
  5. vinnylarouge/jevlike+444★654
  6. NationalSecurityAgency/ghidra+407★78251
  7. addyosmani/agent-skills+386★95231
  8. affaan-m/ECC+378★259461
  9. ctdal/cve-2026-41940-PoC+306★378
  10. korcarc/text-humanizer+289★365