Python Tools

Text-to-CAD is a library of agent skills that enables AI to perform focused engineering tasks. It solves the problem of automating complex CAD, simulation, and fabrication workflows by providing ready-to-use skills for creating models, finding parts, generating robot files, and previewing designs. It's for developers and engineers building AI agents for mechanical design, robotics, and manufacturing.

A skill to stop your coding agent from burying the answer. ADHD-friendly output.

Markitdown is a Python tool that converts various file formats and office documents into Markdown. It solves the problem of content being locked in proprietary or unstructured formats, making it difficult to edit, version, or publish. It's designed for developers, technical writers, and anyone who needs to work with content in a plain-text, portable Markdown format.

Swarms is an end-to-end agent infrastructure platform designed for building production-grade autonomous systems. It provides the frameworks, interfaces, and cloud services needed to create complex multi-agent collaboration systems with advanced communication, memory, and simulation capabilities. It's built for teams and enterprises that need to ship reliable, scalable, and secure AI agent applications.

Open Science is an AI research workbench that integrates coordinator and specialist agents, Python/R execution, and scientific data connectors into a single desktop workspace. It solves the problem of fragmented research tools by keeping questions, code, execution, and traceable artifacts together in persistent projects for reproducible workflows. It's designed for researchers and scientists on macOS, Windows, and Linux who need a model-agnostic, local-first environment for AI-assisted discovery.

The Skills repository is a public catalog that defines and structures developer skills, competencies, and knowledge areas. It solves the problem of inconsistent or ambiguous skill definitions for AI coding assistants, enabling them to better understand and apply the correct techniques and patterns. This resource is primarily for developers, AI engineers, and tool builders who are creating or integrating AI-powered development tools that need a common reference for technical skills.

Miles is an open-source reinforcement learning framework designed to make training large language and vision models simple, efficient, and stable. It solves the engineering complexity of agentic RL for post-training tasks like SFT, RL, and distillation, enabling faster iteration and stable long runs. It's built for AI labs and enterprises working with frontier models, offering proven scale, extensive hardware support, and integrations with popular agentic environments.

This repository is a collection of public exploit Proofs-of-Concept (PoCs) and vulnerability research write-ups for unreported security flaws. It aims to lower the barrier to entry for security research by providing concrete, real-world examples for study and learning. It is intended for aspiring security researchers, bug bounty hunters, and students looking to understand vulnerability discovery and exploitation techniques in a practical, hands-on context.

Humanizer is an agent skill that transforms AI-generated writing by removing robotic patterns and unnatural phrasing. It solves the problem of content that feels artificial, helping writers, marketers, and developers produce text that resonates with human readers. This tool is ideal for anyone using AI assistants who needs polished, natural-sounding output for blogs, emails, or communications.

SIE is an open-source inference server and production cluster that allows developers to run all the models their AI agents need—from embeddings and rerankers to large language models—entirely within their own infrastructure. It solves the problem of vendor lock-in, data privacy concerns, and high costs associated with proprietary model APIs by providing a performant, self-hosted alternative. It's built for teams building production AI agents who require control, data sovereignty, and the ability to compose the best open models.

Sequoia-X is an automated A-share stock screening system. It solves the problem of manually scanning hundreds of stocks for complex technical patterns by running automated scans after market close. It then delivers actionable alerts directly to Feishu (Lark), saving time for retail investors and quantitative traders focused on the Chinese stock market.

TimesFM is a pretrained time-series foundation model from Google Research. It solves the problem of accurate, zero-shot forecasting across diverse domains without task-specific training. It is designed for researchers and practitioners who need robust, general-purpose time-series predictions.

VoiceStudio is an open-source desktop application that provides a fully-local alternative to services like ElevenLabs. It solves the problem of data privacy and vendor lock-in by allowing users to perform professional voice cloning, video dubbing, transcription, and audiobook creation entirely on their own computer. It's designed for content creators, developers, and privacy-conscious users who need powerful AI voice tools without sending their data to the cloud.

Academic Research Skills is a suite of open-source AI tools designed for academic researchers. It solves the problem of navigating the complex research-to-publication process by providing structured guidance through research, writing, peer review simulation, and revision. It's built for students, scholars, and academics who want to enhance their workflow with AI assistance while maintaining research integrity and receiving Socratic coaching.

Minimind is a complete, open-source pipeline for training small, efficient language models from the ground up. It solves the problem of LLM training being prohibitively expensive and complex, offering a full-stack solution from tokenization to advanced RL training. It's for developers, researchers, and students who want to deeply understand and practically build LLMs without massive computational budgets.

这是一个专注于中国专利领域的智能助手，旨在帮助创新者、研发人员和中小企业解决专利挖掘难、交底书撰写繁琐的问题。它能通俗解读专利信息，嗅探政策动向，并辅助进行审查答复。无论是发明、实用新型还是外观设计专利，它都能提供从创意到申请的全流程支持。

Microduck is a 25 cm open-source bipedal robot designed for reinforcement learning. It solves the problem of inaccessible, expensive robotics platforms by providing a ready-to-train, sim-to-real system out of the box. It's for developers, researchers, and hobbyists who want to experiment with and train real-world robot behaviors using RL, from simulation to physical deployment.

Crawl4AI is an open-source web crawler and scraper specifically designed to be friendly with Large Language Models (LLMs). It solves the problem of extracting clean, structured, and context-rich data from websites for AI applications, offering features like adaptive crawling, markdown generation, and anti-bot bypass. It's built for developers, data scientists, and AI researchers who need reliable, scalable web data for training or augmenting their LLMs and AI models.

User-scanner is a comprehensive OSINT suite that extracts deep data from just one email address or username. It solves the problem of fragmented online investigations by analyzing over 455 actively maintained scan vectors, including 175+ for emails and 280+ for usernames. This tool is designed for security researchers, investigators, and professionals conducting digital footprinting and threat intelligence.

ODS transforms your personal computer into a powerful, self-hosted AI platform. It solves the problem of privacy, cost, and latency by bringing LLM inference, chat, voice agents, and image generation directly to your local machine. It's ideal for developers, researchers, and privacy-conscious users who want full control over their AI tools without relying on cloud services.

Heretic is a tool that automatically removes built-in censorship and safety restrictions from language models. It ensures AI assistants follow user instructions precisely, rather than refusing requests based on developer-imposed filters. This is for developers, researchers, and power users who need uncensored, directive-following AI for creative, technical, or analytical tasks.