Ai agent

OpenScience is an AI-powered workbench for scientific research that automates the full research loop—literature review, hypothesis formation, code...

A self-hosted quantitative trading workbench for Chinese A-stock analysis, combining stock screening, monitoring, and backtesting with LLM-powered...

OpenBiliClaw is a local-first, privacy-focused AI agent that learns user preferences through cross-platform activity and feedback, then proactively...

AI Berkshire is a specialized investment research framework that applies value investing methodologies from Warren Buffett, Charlie Munger, Duan...

Deep Code CLI is a terminal AI coding assistant optimized specifically for the DeepSeek-v4 model, supporting advanced reasoning modes, thinking depth...

Stock SDK is a TypeScript/JavaScript library for fetching real-time stock market data (A-shares, Hong Kong, US stocks, and mutual funds) directly in...

Zero is a terminal-based AI coding agent written in Go that lets developers use their choice of LLM models (OpenAI, Anthropic, Gemini, Ollama, etc.)...

TencentDB Agent Memory is a TypeScript library that provides local long-term memory for AI agents through a 4-tier progressive pipeline with symbolic...

Dao Code is an open-source terminal-based AI coding agent optimized for DeepSeek V4, designed to read code, write code, run commands, and fix bugs...

LetsFG is an AI agent-native flight search and booking platform that integrates with 400+ airlines to find and book flights without markup or price...

Gini Agent is a personal AI agent platform designed to maintain memory, learn from interactions, and operate autonomously across multiple interfaces...

DashiAI PPT Skill is an AI-agent-integrated presentation generator that creates browser-editable, multi-theme presentations exportable to HTML, PDF,...

Omnigent is an open-source orchestration layer and meta-harness for managing multiple AI agents (Claude Code, Codex, Cursor, and custom agents) with...

VulnClaw is an AI-driven penetration testing CLI tool that automates the full attack lifecycle—reconnaissance, vulnerability discovery, exploitation,...

open-computer-use is an open-source MCP-wrapped computer automation service that enables AI agents to control macOS, Linux, and Windows systems...

A comprehensive trilingual learning roadmap and curated resource collection for AI agents, spanning from LLM fundamentals to multi-agent systems. It...

A comprehensive, systems-level Chinese-language tutorial for learning AI agents and large language model application development, structured from...

A specialized skill package for Chinese web novel writing that covers market research, story structure analysis, composition, AI-style removal, and...

A specialized Claude Code skill for generating social media carousel templates (Xiaohongshu/Rednote and WeChat cover pairs) from article text,...

Ian Xiaohei Illustrations is a Codex Skill that guides AI agents to generate illustrated article diagrams in Chinese—converting abstract concepts,...

A curated library of 1,150+ AI agent skills for empirical research across 8 social science disciplines (economics, political science, sociology,...

Agent Reach is a Python CLI tool that provides AI agents with unified internet capabilities—reading and searching Twitter, Reddit, YouTube, GitHub,...

This is a learning project that reconstructs Claude Code's core architecture in ~5000 lines of TypeScript/Python, designed to teach developers how AI...

CyberStrike is an open-source AI-powered offensive security agent that automates penetration testing by integrating with LLMs (Claude, GPT, etc.) to...

Oh My Pi is a terminal-based AI coding agent that integrates IDE capabilities, LSP support, and multi-provider LLM access (Claude, OpenAI, etc.) for...

Vibe-Trading is a Python-based personal trading agent platform that integrates LLM capabilities with algorithmic trading, backtesting, and...

DeepSeek++ is a browser extension that transforms the DeepSeek web interface into an AI agent workspace with MCP tool support, long-term memory,...

A specialized AI skill that equips Claude, Codex, Cursor, and other AI agents with user research capabilities—planning studies, conducting interviews...

Mercury is a TypeScript-based AI agent that emphasizes user control through permission-hardened tool execution, token budgets, and a persistent...

An educational architectural deep-dive into Claude Code's internals based on leaked TypeScript source code, documenting the agent loop, context...