59 resources
GNOS is an AI teaching assistant that can customize courses for you like a real teacher. It plans routes based on learning depth and time, generates videos, simulations, images, PDFs and other content, records your learning style and reasoning process, and dynamically adjusts subsequent exercises to achieve a personalized adaptive learning experience.
Digital human video production skill for Codex. Input copywriting and authorized character images, automatically complete dubbing, mouth shape calibration, subtitles and editing rendering, and generate high-quality explanation videos. The source code is neutral and not tied to a service provider. It supports flexible configuration of sound, style and other parameters to achieve automated video production.
This is a collection of skills, workflows and use cases provided for Jev coding agents. It contains 61 project resources, 5 core skills and 108 application scenarios, designed to help agents select, classify, score and perform specific tasks, and provides rich examples and templates to enhance its functionality.
AnyJev is a Python library that uses the hidden states of LLM to build Jev-like decision models. It can output decisions with type and real probability without fine-tuning, significantly improves calibration and resistance to sequence flipping, and is suitable for scenarios that require high-reliability automatic decision-making.
This is an "AI-free" skill based on research on a 2.83 million-word corpus, which aims to quantify and remove specific style features of text generated by large models through linguistic methods. It provides a recalculable judgment program to help users identify and correct AI writing traces and improve the humanization of the text.
Hide product video skills, use AI to reuse real product components and design language, and automatically generate software update promotional videos. It supports storyboard copywriting, original soundtrack, sound effects and code rendering, and is compatible with Claude Code and Codex, allowing developers to quickly produce professional promotional videos.
Jev Ultrafast is an extremely fast Python-based web agent that enables efficient browser automation through dynamically indexed action spaces. It uses TypeSafe LLM to perform clicks, inputs and other operations, significantly reducing latency and costs, and is suitable for scenarios where web page interaction tasks need to be completed quickly.
jev-seo is a Python-based real-time SEO audit tool that supports command line and Claude Code integration. It crawls the home page of the website, checks 52 indicators according to Google search center rules and evaluates core web page indicators, and uses the Jev AI model to analyze the page content. Finally, detailed reports in PDF, Excel, and Markdown formats are generated to help quickly diagnose and optimize website SEO performance.
geo-score is a Python command line tool for evaluating a website's visibility in AI engines. It provides a readiness score of 0-100 points based on the open GEO standard, and supports tracking citations through APIs such as OpenAI and Perplexity. Zero dependency, with MCP server function, helping developers quickly detect and optimize AI accessibility.
Laya-MLX is a native Python library running on Apple Silicon for performing ultra-fast (7-14ms) typed decision-making tasks. It is based on the MLX framework, does not require PyTorch or cloud API, supports offline local reasoning, and is suitable for scenarios that require low-latency structured output.
Rizzo Flow is an open source, local-first LLM decision-making tool compatible with the Jev API. It transforms unstructured state input into typed probabilistic decisions without generating text tokens. Supports the llama.cpp backend and can run in Metal, CUDA, Vulkan, ROCm, SYCL and CPU environments. It is suitable for local AI application development that requires low latency and high privacy.
Dewarr is a self-hosted audiobook discovery and management tool that supports connecting to Audiobookshelf/Grimmory, synchronizing Goodreads and Hardcover book lists, and automatically downloading from sources such as MyAnonamouse. It provides chapter merging, personalized recommendations and multi-platform notification functions to help users centrally manage reading lists and download tasks.
Hermes Jev Skills provides low-cost decision support for AI agents, covering model routing, search filtering, memory compression and skill selection. Use TypeSafe's Jev decision-making model to process non-text decisions, save expensive model tokens, and improve response speed.
T-Mem is a triggered enhanced graph memory framework for long-term dialogue question answering, aiming to solve the problem of limited LLM memory accessibility. It supports descriptive and associative recall and is suitable for building conversational agent systems with coherent contextual understanding and cross-session adaptability.
This tool automatically identifies repeated requests and generates shortcut commands by analyzing the user's historical prompts in Claude Code, Codex or Cursor. It is designed to reduce repeated input, improve coding efficiency, and support the rapid reuse of common instructions in mainstream AI programming agents.
/brag is an AI programming assistant skill that automatically converts projects into short videos containing music, motion effects, and sharing copy through a single command. Supports mainstream agents such as Claude Code, Cursor, and Codex to help developers quickly display results.
DocJev is a document classification and splitting tool based on the Jev engine, supporting PDF, DOCX and other formats. Automatically identify document categories or boundaries through natural language rules, and provide CLI and Python library interfaces, which are suitable for efficient parsing and management of batch processing of complex document packages.
RUCareer is a Windows local experience management and AI resume generation tool for college students. It archives users' real experiences into local files, supports the generation of personalized resumes through the OpenAI API, and provides DOCX/PDF export and automatic backup functions, emphasizing data privacy and authenticity.
Laya is a multilingual, non-autoregressive System 1 decision engine that supports typed decision making in more than 100 languages. It is trained through reinforcement learning to achieve a fast response of 33 milliseconds in a single forward pass, and has a router function that automatically selects checkpoints based on requests. It is suitable for scenarios that require low latency and high concurrency decision-making.
Niimbot B1 label printer firmware patch removes the original printing brightness limit for third-party consumables, making user density settings truly effective. Only applicable to version 5.22 firmware, flashed through USB serial port, designed to fix hardware defects and improve print quality.
BrainAPI is an AI memory layer based on knowledge graphs, built using Python. It can automatically convert unstructured text into a structured knowledge graph, support intelligent search, recommendation and contextual memory functions, and is suitable for providing traceable semantic reasoning capabilities for AI agents and applications.
Claude Unlimited is a local CLI tool that seamlessly switches accounts in Claude Code to break usage limits by rotating Claude, ChatGPT and API keys. It supports multi-account pooling, automatically switches between requests to ensure uninterrupted sessions, and provides local dashboard management.
This is an experimental controller that allows TypeSafe's Jev model to directly select the control input of NES "Super Mario Bros." through structured JSON instead of screenshots. It parses simulator telemetry data, automatically performs game operations, and is used to explore AI's decision-making capabilities in structured states.
Doubao Say is a Linux voice input application based on GTK4, supporting Wayland and X11. It uses Beanbao or Volcano engine API to implement speech-to-text, provides multi-language settings, microphone selection and custom trigger shortcut key functions, aiming to bring a convenient voice interaction experience to Linux users.
Artemis is a Python-based Android automation framework that converts natural language instructions into reliable end-to-end workflows. It supports multi-modal AI assistant integration, has cross-application automation, element positioning and IDE diagnostic functions, and achieved a high success rate of 99%+ in the AndroidWorld benchmark test.
APXinf-robo is a high-performance edge inference engine developed based on Rust, specifically designed for embodied AI. It supports VLA/WAM models (such as PI-0.5) on Jetson Thor/Orin devices, providing BF16/FP8/INT8 accuracy optimization. Simplify custom model development through Agentic workflow, suitable for robot strategy execution and benchmark testing.
page-mascot is a React component for embedding interactive mascot characters in web pages. It can follow the movement of the mouse cursor and blink when clicked. It supports custom image drawing and multiple attribute configurations, aiming to add a vivid visual interactive experience to web applications.
Conflux is a modern, neutral and beautiful Linux icon theme designed to make all icons consistent in style. It features a free-form design, is compatible with unthemed icons, and retains the original high-quality icons. Achieve visual integration by modifying the background or adjusting the position, helping users create a seamless and natural desktop environment.
Droid ASC is a super-fast Android decompiler front-end for mobile security researchers. It achieves millisecond-level code search and extraction by directly querying compiled APK files instead of building heavy indexes. This tool takes advantage of the R8 compiler optimization features to skip traditional preprocessing steps, significantly reducing memory usage and improving analysis efficiency.
Neutrino is a Python-based developer center for managing network access, credentials, machines, AI and remote workflows. The Hub runs on the Linux host, the Agent is deployed on the managed Linux machine, and the Client supports Linux, Windows and macOS, realizing unified access to services on the LAN or external network.
ThinkingBox is an LLM agent framework launched by Microsoft, which is used to define MCP tools, run agent interactions and evaluate their behavior. Supports offline training data generation and reinforcement learning loops. The main functions include creating scenarios, simulating user-agent interaction and result evaluation, and is suitable for model training and evaluation.
Miles is an enterprise-oriented reinforcement learning framework designed for post-training of LLM and VLM. It supports efficient fine-tuning and optimization of multi-modal models and provides enterprise-level solutions designed to improve the performance and productivity of large-scale language and visual models.
WhisperLive is a real-time speech-to-text tool based on OpenAI Whisper, which supports microphone live broadcast and audio file transcription. It provides command line interface, Docker deployment and browser extensions. It has advanced functions such as word-level timestamps and speaker separation, and is suitable for multi-platform server-side deployment.
Maskit is a local privacy desensitization gateway specially designed for AI programming assistants. It automatically performs structured coding on sensitive information (such as keys, intranet IP, and PII) before sending a request, and streams it back when replying. Supports tools such as Cursor to ensure 100% local processing of data with zero telemetry, ensuring development privacy and security.
This warehouse provides configurations, benchmarks, and ARM64 runtime images for reproducible large language model inference on a four-node NVIDIA DGX Spark cluster through a switchless direct connection. It is designed to demonstrate high-performance deployment solutions under specific hardware topologies, including detailed performance data and verification results.
UMR is a unified action redirection code library for humanoid robots. It maps motion data from humans or other sources to the robot by learning point cloud correspondences. It mainly includes two stages: point cloud correspondence learning and correspondence-based motion optimization, supporting efficient conversion and adaptation of multiple motion source data such as bones and surfaces.
PivotHub is a Python-based CTF multi-layer intranet penetration aid. Using FastAPI backend and Vue 3 frontend, it provides intranet topology visualization, command library management and shooting range drill functions through the Web interface. Designed specifically for authorized environments, it supports localized deployment and SQLite data storage, and is suitable for teaching demonstrations and competition scenarios.
Misaki is a G2P (text-to-phone) engine designed specifically for the Kokoro model, supporting English and Japanese processing. It provides a Python library interface that can generate phoneme sequences and token mappings, and supports rollback processing through espeak, which is suitable for text preprocessing tasks before speech synthesis.
Convert PDFs, books and other materials into traceable and testable interactive learning web pages. Supports offline operations, emphasizes the separation of facts and model content, and is suitable for AI Agent environments with file access and Python execution capabilities.
CAD-MCP is a CAD control service based on Python that drives software such as AutoCAD, GstarCAD and ZWCAD for drawing through natural language instructions. It supports line, graphic drawing and layer management to achieve CAD automation without manual operations. It is mainly targeted at Windows platform users.
Provides an AI skills package for building Three.js browser games, integrating game development, UI, asset generation and automated testing. The director mode automatically coordinates various specialized skills, supports AAA-level image quality and complete workflow, and eliminates the need for manual selection of tools. It is suitable for quickly developing playable, high-quality web games.
SmartPhotoCrafter is an end-to-end automated photographic image editing framework. It automatically generates image quality analysis and performs repair (denoising, enhancement) or retouching (tonal adjustment) through closed-loop reasoning without user instructions. The tool supports high-fidelity generation and is designed to enable fully automated photo optimization and beautification.
Vibe Squad is a Python-based multi-model AI orchestration tool that defines behaviors through Markdown. It uses a coordinator to distribute tasks to 71 professional roles, execute them in parallel in isolated git workspaces, and supports independent reviews across model families (such as Claude, Codex, etc.), aiming to provide an efficient and safe automated workflow for structured coding.
TempAct is a reinforcement learning framework for autoregressive video generation, aiming to improve the rationality of action timing in videos. It jointly optimizes the LLM planner and diffusion executor to ensure that the generated video strictly follows the time sequence in the prompt words (such as pouring the sauce first and then adding the cheese), solving the problem of timing confusion in the traditional method, and is suitable for video generation scenarios that require high-precision event progress control.
ffmpeg-skill is a FFmpeg skill package for AI encoding agents (such as Claude Code, Cursor). It provides 40 tools to support media editing tasks such as video editing, audio processing, subtitle adding, and format conversion. No cloud or API Key required, purely local operation, suitable for Python 3.9+ environment.
invest-skills is a research assistant for A-shares that automatically collects public data, calculates indicators and generates reviewable research memos through the Python engine. It emphasizes data source tracing and cross-validation, aims to assist learning and research, is not an investment decision-making tool, and ensures that every number can be traced back to the original source.
This is an AI agent operating system for Blender 5.2, which ensures the generation of 3D printable geometric models through bpy knowledge verification and production access control mechanisms. Supports automated modeling of agents such as Claude Code to achieve a closed loop from design to manufacturing, and is suitable for macOS and Linux platforms.
L0p4Map is a professional network monitoring and visualization tool that combines Nmap scanning, ARP discovery and real-time topology engine. Designed specifically for security researchers, it provides continuous monitoring, device fingerprinting and abnormal alarm functions to help users quickly grasp the status of network infrastructure.
Based on Python's Codex Skill, convert manuscripts or SRT into stroke-by-stroke hand-drawn explanation videos. It supports single scenes, multi-act stories and dual semantic island structures, provides Q version characters and Xiaohei style, and outputs real MP4, subtitles and editable materials to achieve the effect of speaking and drawing at the same time.