Custom agents

ytopt: machine-learning-based autotuning and hyperparameter optimization framework using Bayesian Optimization

Tidy is an open-source agentic development environment for engineers and designers. Apache-2.0.

Hora kit core modules

All-in-one desktop toolkit for NVIDIA Jetson: one-click JetPack/L4T firmware flashing, SSH remote development, device diagnostics, app market (YOLO, LLM, RAG, robotics, ROS), AI agent skills (Claude/Codex), OTA updates and disk backup. PyQt6 GUI for Linux & Windows

CloakBrowser Github: anti-detect browser download, source-level chromium patches, browser fingerprinting spoofing, multi-accounting tool, canvas webgl modification, proxy manager http socks5, cookie isolation, session import export, user-agent switcher, puppeteer playwright automation api, profile sharing, web scraping free, github

geminese is an Obsidian plugin that embeds Gemini CLI as an AI collaborator in your vault. Your vault becomes Gemini's working directory, giving it full agentic capabilities: file read/write, search, bash commands, and multi-step workflows.

A collection of OpenClaw Agent Skills — search, analysis, content extraction, and more.

Security-hardened openclaw with auth gateway, AES-256 encryption and session management

Cookbooks and tutorials on Literal AI

XCoding: a lightweight AI vibe coding IDE that supports parallel multi-agent collaboration (Codex/Claude Code) and parallel multi-project development, featuring a VS Code–like editor and terminal with task-driven workflows and code change management.

Extensible ChatOps Framework to enable other chatbots and the Nautobot chat command.

An example notebook to build RAG with Excel files using SQL Agents.

EnvoyJS is a simple JavaScript framework for building AI agents.

A framework for Telegram Bot API written in PHP.

Host your own LLM ChatBot!

Crawl a website to generate knowledge file for RAG

Your AI Powered Enterprise Knowledge Partner. Designed to be used at scale from ingesting large amounts of documents formats such as pdfs, docx, xlsx, png, jpgs, tiff, mp3, mp4, jpeg. Integrates with s3, Windows Shares, Google Drive and more.

Kubernetes operator for Qdrant

Introducing DataKitsune, an open-source Telegram bot designed to enhance the way you manage and retrieve links shared within your personal or group chats. By automatically indexing the content of these links, Data Kitsune allows you to efficiently search and access shared resources based on their content.​

🌲 Code for our EMNLP 2023 paper - 🎄 "Tree of Clarifications: Answering Ambiguous Questions with Retrieval-Augmented Large Language Models"

GraTAG — Production AI Search via Graph-Based Query Decomposition and Triplet-Aligned Generation with Rich Multimodal Representations

Multi-agent reinforcement learning for autonomous navigation for mapping and multi-objective drone swarm exploration

NanoCoder Pro — Autonomous Coding Agent with Master-SubAgent Architecture

Ricochet is the first hybrid coding agent that lives in your IDE and your pocket. Code autonomously with a powerful VS Code assistant, then switch to Ether Mode to control your development environment remotely via messenger. Build, debug, and deploy from anywhere.

AgentPay is a decentralized marketplace where autonomous AI agents provide specialized services (DeFi optimization, smart contract auditing, creative writing) in exchange for MNEE Stablecoin payments. This project demonstrates Programmable Money by combining Gemini 2.5 Flash AI with simulated MNEE blockchain settlements.

A leader-follower formation control using deep reinforcement learning environment, In which every agent can learn to follow the leader agent by keeping track of a certain distance to that leader, avoiding obstacles, and avoiding collision with the other agents.

Codes for paper of 'Solving job scheduling problems in a resource preemption environment with multi-agent reinforcement learning'

Agent exposing connection-related information, based on the Elastic Beats framework

Shell copilot - sh shell AI copilot

ChatBot, show how to implement a RAG based on OceanBase or OceanBase seekdb AI capabilities escpecailly hybrid search and AI embedding.

Optimize AI workflows with Arachne. Automatically assembles the perfect code context (Tree, Target, Deps, Semantic) to fit context windows without noise. Built for efficiency and scale.

A universal orchestration workbench for macOS. Seamlessly mix AI models (Codex, Claude) with traditional CLI tools (git, npm) via static DAG pipelines or dynamic self-healing agents, creating a fully automated closed-loop for development and ops.

A lightweight dependency-free workflow automation platform. Supports iPaaS, stream computing, MCP, and AI capabilities.

A MCP (Model Context Protocol) server that provides get, send Gmails without local credential or token setup.

LLM using long-term memory through vector database

Autonomous driving episode generation for the Carla simulator in a gym environment. This framework makes it easy to create driving scenarios to train/test the agent.

[ICLR 2026] M2-Miner: Multi-Agent Enhanced MCTS for Mobile GUI Agent Data Mining

An intelligent task offloading method based on multi-agent deep reinforcement learning in ultra-dense heterogeneous network with mobile edge computing

LiteMind 是基于 Java 21 + Spring Boot 3 + Spring AI + RAG + Tool Calling + MCP 构建的通用 AI 智能体。支持多轮对话、记忆持久化和 RAG 知识库检索，基于 ReAct 智能体工作模式，具备自主思考能力，能够调用工具完成复杂任务，如网页搜索、资源下载、PDF 生成等。

.NET implementation of the Agent2Agent (A2A) protocol to enable secure, interoperable communication between autonomous agents across frameworks and vendors.

A production-grade multi-agent system for comprehensive medical diagnosis and coding using specialized AI agents.

My first attempt at building an AI "agent". The agent will try iteratively solving Advent of Code puzzles.

Create a chatbot in Microsoft Word powered by ChatGPT

A highly customizable chat Web Component that integrates easily with shadcn - works with any framework (React, Vue, Angular, Svelte, etc.)

Enhanced and elegant flexible peak/loop/domain -calling and analysis tool for 1D/3D genomic data.

AgentPro is a lightweight library for developing ReAct-style agents with tools, knowledge base, memory and reasoning.

Code for NeurIPS 2024 paper "AutoManual: Constructing Instruction Manuals by LLM Agents via Interactive Environmental Learning"

The Claude subagents and example custom commands I use in my AI-assisted development workflow

Claude Code Agent Team GUI Monitor - Tauri v2 + React + TypeScript macOS desktop app with pixel art office view

An agentic AI for tech career coaching 程序员技术学习AI智能体 (based on JoyAgent-JDGenie with local RAG and code analyzing)

MASA: Developing A Multi-Agent and Self-Adaptive Framework with Deep Reinforcement Learning for Dynamic Portfolio Risk Management.

[ToN 2024] Ensuring Threshold AoI for UAV-assisted Mobile Crowdsensing by Multi-Agent Deep Reinforcement Learning with Transformer

Agent Skills Evaluation Framework

A 5-layer asynchronous AI agent framework with Docker-in-Docker sandboxing and hybrid memory.

LAMARL: LLM-Aided Multi-Agent Reinforcement Learning for Cooperative Policy Generation

A multi-agent research MCP server + mini client adapter - orchestrates a net of async agents or streaming swarm to conduct ensemble consensus-backed research. Each task builds its own indexed pglite database on the fly in web assembly. Includes semantic + hybrid search, SQL execution, semaphores, prompts/resources and more

Official Code for "Continuous Locomotive Crowd Behavior Generation (CVPR 2025)"

Official Implementation of AAAI 2025 paper "MDD-5k: A New Diagnostic Conversation Dataset for Mental Disorders Synthesized via Neuro-Symbolic LLM Agents"

Multi-agent system that designs and generates complete agent team structures through conversational interviews. Supports Claude Code and Codex.

AI Agent技术白皮书（Google，2025）中文翻译

Outline-Driven Development - opinionated methodology between vibes and specs. AST tooling + context engineering + loop engineering on your hands. Plugins for Claude / Codex / Gemini / Cursor.

It is a simple, fast, and hard-durable embedded database designed specifically for AI agent memory. It provides a single-file-like experience (no server required) but with native support for vectors, graphs, and temporal search.

DeFi Trading Agent MCP Server - Transform your AI into an autonomous crypto trading agent

A hybrid retrieval system for RAG that combines vector search and graph search, integrating unstructured and structured data. It retrieves context using embeddings and a knowledge graph, then passes it to an LLM for generating accurate responses.

An AI benchmark for Pokémon VGC with agent implementations using multi-agent reinforcement learning, behavior cloning, LLMs, and heuristics

Privacy-first semantic code navigation MCP server using RAG. Features deep AST multi-language support (Go, PHP/Laravel/WP, JS/TS/React, Python), 100% local LLMs (Ollama), and vector search (Qdrant) for AI IDEs like Antigravity, Cursor, Windsurf, Copilot, and Claude.

memweave is a zero-infrastructure, async-first Python library that gives AI agents persistent, searchable memory — stored as plain Markdown files

Self-improving behavioral files for AI coding agents. Analyzes session transcripts for correction patterns and makes surgical edits to prevent recurrence.

An opinionated full-stack agent harness with native memories, long running tasks, and multi-agent relay

On-demand design systems from any URL. Five surfaces, one agent: web, API, CLI, TypeScript SDK, and a portable agent skill that runs inside Claude Code, Codex, and Cursor.

Markdown-first Notion MCP server. ~6-7x fewer response tokens vs official Notion MCP. 43 tools.

OSINT intelligence MCP server for AI agents — 37 tools, 12 sources. Shodan, VirusTotal, Censys, SecurityTrails, DNS reconnaissance, WHOIS, certificate transparency, BGP routing, Wayback Machine, GeoIP. Automated open source intelligence and attack surface mapping via Model Context Protocol.

VeriFlow-CC: A Claude Code-driven RTL design pipeline. Automates Chip-on-Chat from architecture to synthesis (iVerilog/Yosys) using a stateful, zero-dependency LLM orchestration skill. Features sub-agent nesting for code gen and behavioral-driven verification.

🤖 LLM Engineering Roadmap — Complete Developer Guide

Multi-agent AI framework that writes full-length novels with persistent story state, a deterministic continuity engine, and a five-role editorial pipeline.

Free AI models in VS Code Copilot Chat / DSH via browser auth. No API keys required.

An AI coding harness that shaped itself - repo awareness, surgical edits, validation, recovery, and safe diff approvals.

Connects any MCP-compatible AI assistant to Siemens TIA Portal via the Openness API

Python CLI tool to run queries against sqltatabases and convert the results to json, csv, excel in the command line or python program. Easy to use for Humans, automations, LLMS/Agents

[NeurIPS 2024 D&B] VideoGUI: A Benchmark for GUI Automation from Instructional Videos

Airgapped closed-corpus QA loop: a self-hosted Qwen3.6 agent explores a .zip dataroom under a token budget with local tools

SCALE Engine 是一个 AI 工程化脚手架引擎，为 AI Agent（Claude Code、Codex CLI、OpenCode、Cursor、Gemini CLI 等）提供物理约束层，确保 AI 在编码过程中遵循工程规范，而不是依赖提示词的"自律"。

A harness framework for building domain-specific agents, the worked finance instance that beats every frontier model on public benchmarks, and the autonomous loop coordinator that drives 100-experiment optimization runs unattended.

See which AI coding agent needs you — right in your tmux tab bar. A colored dot per window (working/blocked/idle) with off-screen sound alerts. Works with opencode, pi, Claude Code & Codex.

Git for Agent Memory: Copy-On-Write vector branching for embedded multi-agent memory (83x faster, 3000x smaller snapshots)

VoidClaw is a universal autonomous AI command centre that runs seamlessly on Windows, Mac, Linux, and Android. It acts as a local-first digital assistant that manages your files, automates web tasks, and controls your devices with an evolutionary memory that adapts to your unique workflow.

"hamroh" is a AI agent harness framework for running your own persistent AI companion on Telegram — one you fully own, control, can extend, one that learns from you. Agent template: https://github.com/Rustam-Z/hamroh-custom-agent

Official benchmark reproduction code for the SAG paper

Build an agent harness from scratch. 21 runnable Python lessons, 3 reliability design labs, 78 offline tests, no agent frameworks.

Neuroscience-inspired memory framework for AI agents

AI浏览器 MCP Server — Windows 本地浏览器自动化 MCP 服务端 · 334 个工具全量真机核验(0 失败/0 闪退) · Cursor/Claude Code/Claude Desktop/Trae 原生 stdio 直连 · Web Scraping · JS 逆向 · CDP 断点 · 指纹反检测 · 网络拦截 · 填表 RPA · FBrowser CEF 内核 · MIT 开源

A Survey for LLM Agent Trajectory Analysis: From Failure Attribution to Enhancement(Accepted by TSE 2026)

GitHub Copilot modernization agent

sofagent is an open-source FDE Harness for AI coding agents — on entry, write your business judgment into files (workflows, ontology, AI nodes); after departure, audit every change against them. 28 codified audit rules, tamper-evident chain, snapshots, and an MCP toolset (104 tools, 12 plugins). Eve

Open-source alternative to Maestri. Native macOS multi-agent orchestration canvas — connect, coordinate & deploy Claude Code, Codex, OpenCode, Gemini CLI agents visually.

Ready-to-use AI agent skills for e-commerce intelligence. Product research, keyword optimization, competitor analysis, pricing monitoring, patent risk assessment, market trends, and platform-specific tools for Amazon, TikTok Shop, Shopify, Temu and more.

The curated list of tools, standards, and resources for governing AI agents in production: policy enforcement, audit trails, EU AI Act compliance, MCP and Claude Code security.

Podlite markup language: parser, CLI (convert, lint, query, test), and MCP server for AI agents

Run large models like Qwen3.6-35B-A3B locally on a 16 GB RAM machine