Agent

Mano-P is a GUI-VLA agent project designed to enable autonomous, private task execution on edge devices like Mac mini and MacBook. It utilizes advanced reinforcement learning and edge-native inference to perform complex GUI automation, cross-system data integration, and long-task planning. The project provides a secure, local-first solution that eliminates the need for cloud API calls while maintaining high performance across various benchmarks.

DeerFlow is an open-source super agent harness designed to orchestrate sub-agents, memory, and sandboxes for complex task execution. The platform features a ground-up rewrite in version 2.0, offering enhanced extensibility through a modular skill and tool architecture. It supports diverse deployment environments, including local development and Docker-based production setups, with integrated support for multiple messaging channels.

OpenSandbox is a versatile sandbox platform designed for AI applications, supporting diverse runtimes like Docker and Kubernetes. It provides multi-language SDKs and a unified API to facilitate tasks such as code execution, agent evaluation, and browser automation. The platform ensures secure isolation through container runtimes while offering robust network controls and lifecycle management.

DeerFlow 2.0 is a ground-up rewrite of an open-source super agent harness designed to orchestrate sub-agents, memory, and sandboxes. It utilizes extensible skills and integrates with various AI models to perform complex tasks through a flexible, containerized architecture. The framework supports multiple deployment modes and provides seamless connectivity with messaging platforms like Slack, Telegram, and Feishu.

WeKnora is an open-source, LLM-powered framework designed for enterprise-grade document understanding, semantic retrieval, and autonomous reasoning. It features a ReAct agent for complex multi-step tasks and a Wiki mode that distills raw documents into a structured, interlinked knowledge base. The platform supports multi-source data ingestion, various LLM integrations, and flexible deployment options to ensure complete data sovereignty.

OpenClaude is an open-source coding-agent CLI that supports a wide range of cloud and local model providers. It offers a unified terminal-first workflow featuring tools for file management, bash execution, and agentic tasks. Users can easily integrate various backends, including OpenAI, Ollama, and Gemini, while leveraging advanced features like agent routing and gRPC support.

Claude Code Game Studios transforms a standard AI coding session into a structured, professional game development environment. It utilizes a hierarchy of 49 specialized agents, 72 workflow skills, and automated validation hooks to maintain project organization and quality. The system ensures developers remain in control while benefiting from expert-level guidance across design, programming, and production phases.

Toonflow-app is an AI workbench designed for short drama production, achieving full-process automation from script to video through an infinite canvas and a three-layer Agent collaboration system. The platform supports chapter event graph-driven adaptation and provides a programmable provider system to flexibly integrate various AI models. Users can leverage its persistent memory system and modular skill configuration to significantly improve the efficiency and consistency of short drama creation.

Hermes Agent is a self-improving AI assistant designed by Nous Research that creates and refines skills through a built-in learning loop. It supports a wide range of LLM providers and can be deployed across various platforms including Telegram, Discord, and local terminal environments. The system features persistent memory, scheduled automations, and the ability to spawn subagents for complex, parallelized tasks.

The OpenAI Agents SDK is a lightweight framework designed for building complex multi-agent workflows. It supports a wide range of LLMs and provides essential features like tool integration, guardrails, and human-in-the-loop capabilities. Developers can also utilize sandbox agents for long-running tasks and leverage built-in tracing to debug and optimize their agentic applications.

EvoCUA is a high-performance open-source multimodal model designed for end-to-end computer automation across various desktop applications. It currently holds the top ranking on the OSWorld benchmark and demonstrates superior cross-OS generalization capabilities. Additionally, the model is recognized for its robust safety profile, exhibiting the lowest unintended-behavior rate among leading computer-use agents.

OpenOcta is an open-source enterprise-grade agent designed for operations, observability, and automation scenarios. The project utilizes a single Go binary architecture, simplifying deployment and distribution through embedded frontend technology. It is compatible with the OpenClaw protocol, enabling unified access to data sources such as monitoring, logs, and CMDB to assist enterprises in troubleshooting and decision analysis.

ZhangXuefeng.skill is a cognitive operating system built on deep research, designed to provide an executable thinking framework rather than a simple collection of quotes. By distilling core mental models, decision heuristics, and communication DNA, the project helps users analyze major selection and career planning from Zhang Xuefeng's perspective. Users can install this skill to obtain targeted decision-making advice and in-depth analysis within Claude Code.

This is a comprehensive and in-depth practical tutorial for OpenClaw, designed to help users build an efficient AI work assistant from scratch. The tutorial covers full-spectrum guidance from installation and deployment to core feature usage and advanced skill expansion, providing numerous ready-to-use practical cases. The project not only supports multiple deployment methods but also deeply integrates with mainstream domestic office platforms, helping individuals achieve a tenfold increase in efficiency.

This tutorial provides a comprehensive guide for users to build an AI work assistant from scratch, covering installation, configuration, core features, and advanced techniques. The content is proofread based on the stable OpenClaw v2026.4.14 version and offers multiple deployment options to meet various scenario requirements. Through rich practical cases and detailed command cheat sheets, it helps users achieve a significant boost in personal efficiency.

EvoCUA is a high-performance, open-source multimodal model designed for end-to-end computer automation tasks. It currently holds the top position among open-source models on the OSWorld benchmark and demonstrates strong zero-shot cross-OS generalization capabilities. Additionally, independent research identifies EvoCUA-32B as the safest computer-use agent, exhibiting the lowest rate of unintended behaviors.

Browser Harness provides a direct connection between LLMs and your browser using a thin, editable CDP interface. The system allows agents to write and improve their own helper functions during execution to handle complex tasks. Users can leverage this framework to automate browser workflows while building a library of reusable, agent-generated domain skills.

Awesome-finance-skills is a plug-in skill collection that provides financial analysis capabilities for large language models. It supports various professional financial functions such as real-time news aggregation, stock data queries, sentiment analysis, and market forecasting. Users can integrate these skills into mainstream AI Agent frameworks through simple installation to quickly enhance their financial analysis level.

Awesome Finance Skills is a plug-and-play collection of financial skills designed to empower AI Agents with real-time news analysis, stock data processing, and market forecasting capabilities. The project supports various mainstream Agent frameworks, allowing users to quickly integrate relevant features through one-click installation or manual deployment. By integrating the Kronos model with multiple financial data sources, it helps Agents automatically generate professional market analysis reports and logical chain diagrams.

Vibe-Trading is an AI-powered multi-agent workspace that translates natural language requests into executable trading strategies and portfolio analysis. It features 71 specialized finance skills and 29 pre-built swarm workflows to automate research, backtesting, and risk management across global markets. Users can easily export their generated strategies to platforms like TradingView, TDX, and MetaTrader 5 with a single command.

dot-skill is a versatile AI framework that distills individuals into interactive digital skills by analyzing their unique thought patterns and communication styles. The platform supports three distinct character families, including professional colleagues, personal relationships, and public figures. It integrates seamlessly with multiple AI agent hosts to provide a unified, automated experience for creating and invoking personalized AI personas.

Claudian is an Obsidian plugin that integrates AI coding agents like Claude Code and Codex directly into your vault. It transforms your vault into an active working directory where agents can read, write, search, and execute bash commands. Users can interact with these agents through a chat sidebar, inline editing, and support for Model Context Protocol servers.

DeepTutor is an agent-native platform designed to provide personalized, intelligent tutoring through a unified chat workspace and multi-agent architecture. It features advanced capabilities like a Book Engine for interactive learning, an AI Co-Writer, and persistent memory to tailor the experience to individual user profiles. Users can deploy the system easily via a guided CLI setup or Docker, supporting a wide range of LLM and embedding providers.

jcode is a high-performance coding agent harness designed for multi-session workflows and extreme resource efficiency. It features a sophisticated memory system that uses semantic vector embeddings to recall relevant information without excessive token usage. The platform supports native agent collaboration through a swarm architecture and integrates with a wide range of LLM providers via OAuth or custom configurations.

The AI Hedge Fund is an educational proof-of-concept project designed to explore how artificial intelligence can be utilized for making trading decisions. It employs a multi-agent system that simulates various famous investment strategies and analytical approaches to evaluate stocks. The system is strictly for research purposes and does not execute real-world financial trades.

AI Engineering from Scratch is a comprehensive 320-hour curriculum that guides students from fundamental linear algebra to building autonomous agent swarms. The course emphasizes an AI-native learning approach where students use AI coding agents to test their knowledge and build reusable tools throughout 20 distinct phases. By working across Python, TypeScript, Rust, and Julia, learners develop a professional portfolio of prompts, skills, and agents that can be deployed in real-world environments.

This repository provides a comprehensive educational framework for building agent harnesses, which are the essential environments that allow AI models to perceive and act. It argues that true agency is learned by models during training, while the developer's role is to construct the tools, knowledge, and context management systems that enable these models to function. Through twelve progressive sessions, users learn to build robust, scalable agent architectures by reverse-engineering the principles behind Claude Code.

Claude How To provides a structured, visual learning path to help developers master the full capabilities of Claude Code. The guide includes ten tutorial modules, copy-paste templates, and interactive quizzes to bridge the gap between basic usage and advanced automation. It enables users to build complex workflows by combining slash commands, memory, subagents, and MCP servers effectively.