Custom agents

Curated map of 3,000+ open-source AI agent tools: coding agents, agent frameworks, MCP servers, RAG, memory and local LLMs, sorted by layer and ranked by GitHub stars.

沐问 MuAsk｜自然语言一键查数据库，AI 自动读取真实表结构生成可执行 SQL，支持 MySQL/Oracle/达梦/金仓等 15 种数据库，驱动全部内置

A modular local AI agent built with Python and Ollama, featuring tool calling, memory, web search, document processing, embeddings, and RAG.

Four-source migration wizard for DeepSeek Harness: move Claude Code, Codex, OpenCode and Hermes sessions, memories, skills, instructions and slash commands into DSH, through a /move wizard with resumable, approval-gated, idempotent sessions. Maintenance status: FROZEN since 2026-10-05.

Claude Code /rewind for DeepSeek Harness — git-first workspace snapshots before every mutation, turn-boundary session forks, one-shot /rewind restore. A dsh-plugin capability seam.

跟随 Codex / Claude Code / Gemini 等 12 个 Agent 实时状态换表情的 Windows Live2D 桌面宠物（只读本地、不联网）

On-device AI agent for Android, tuned for free OpenRouter/Kilo models with automatic model/key fallback and rate-limit recovery. Tools: web search/fetch over HTTP, bash/Python in Termux (optional headless Chromium + Selenium), Shizuku/rish shell, Accessibility, ebook translation, Documents/Verdroid, Git, crypto market data. No backend.

Let Claude and Codex call each other from their desktop apps to review, cross-check or implement, with a live side panel. / 让 Claude 与 Codex 在各自桌面版里互相审查、协作的插件。

Static analysis and attack-surface scanner for AI agents — SAST for prompts, tools, MCP servers, and agent workflows. Detects capabilities, prompt risks, tool permissions, and governance gaps before deployment.

PMX Canvas is a shared board where you and your AI agents collect and present the data and evidence behind your work. Curate that material into context for your agent's next step, then bring the results back onto the board to review and build on. You both add and edit material on the same board, in real time. Pin context to focus input and output.

Meet your crew. APX is an open-source Agent OS for running AI agents on your own machine — memory in plain files, agents that work together, Claude Code, Codex, Telegram, WhatsApp and MCP.

Autonomous agent that mines Java deserialization gadget chains from any JAR directory. Static bytecode analysis + dynamic JVM probes + dual-model adversarial auditing → weaponized PoC. Solving the last mile.

Build smarter. Work superior. An opinionated rule set for AI coding agents to create original interfaces, compelling content, and production-ready code.

Everything is a Plugin · 一切皆插件｜Windows AI 桌面宿主：任务栏／灵动岛、Agent 会话与审批、邮件、独立插件中心、整仓备份及完整／精简双包更新。基于 Wakuwaku，Rust + Tauri。

Let an AI companion get sleepy, fall asleep, dream, and be woken by you. 让 AI 伴侣会困、会睡、会做梦，也会被你叫醒。

MASRubric: Auditing Information Flow in Multi-Agent Systems with Failure-Distilled Pitfall Rubrics.

Desktop status hub & notifier for AI coding agents — know when Claude Code, Codex CLI, OpenCode, Grok & Kimi Code finish or need you, with token usage and API-equivalent cost per tool, project and model.

One terminal for all your AI coding agents. Run Claude Code and Codex side by side, see which one needs you, with Git and a shell one key away.

⌘-click an element in your running app, say what to change, and Claude, Codex, Cursor or another AI agent opens with the exact file and lines. Vite, webpack, rspack, Rsbuild, Next.js and more.

Context window optimization for AI coding agents. Sandboxes tool output (98% reduction), persists session memory via SQLite + FTS5, and enforces routing across 17 platforms through MCP + hooks. Works with Claude Code, Gemini CLI, VS Code Copilot, OpenCode, Codex CLI.

[ICML 2026]A defense framework against MLLM-based web GUI agents. This repository provides both the generative CAPTCHA system and tools for evaluating agent resistance.

An LLM-driven fuzzing pipeline powered by the GitHub Security Lab Taskflow Agent

Agent skills that turn docs/ into an LLM-maintained wiki and task tracker. Obsidian-friendly. Works with Claude Code, Codex and Copilot CLI.

MCP Server for iOS UI inspection — lets any AI agent (Claude Code, Cursor, Codex, etc.) inspect live iOS app UI hierarchy, attributes, and screenshots via LookinServer, no Lookin.app required.

柚子 AI · A 股 AI 研究项目包，公众号「AI短线引擎」：行情查询、AI 盯盘、AI 盘面分析、AI 复盘、模拟交易 Skill、定时调度一体化 SDK | 适配 Claude Code / OpenCode / Codex / Workbuddy / ZCode / KimiWork 等 AI Agent 平台

MCP server for managing XBloom Studio coffee recipes with Claude

Cross-harness coding orchestration for Claude Code. Delegate to Codex, Antigravity, or custom CLI agents via pluggable adapters.

Reusable Codex/Claude skills and local agent configs

Keenetic router control for Claude Code, Codex, Cursor and any MCP agent. An MCP server over the router's own RCI API, plus ready-made plugins whose skills teach the agent how the hardware really behaves. Devices, Wi-Fi, VPN routing, isolated segments. Nothing installed on the router.

Long-Horizon Memory-Augmented LLM Agents

AI/LLM on Kubernetes 基础设施知识库：涵盖 GPU、大数据、调度、分布式训练、LLM 推理、RAG、Agent 与生产运维。| AI/LLM on Kubernetes Infrastructure Knowledge Base: Covering GPUs, big data, scheduling, distributed training, LLM inference, RAG, agents, and production operations.

考研管综199历年真题题库(2009–2026)：57题/年·试卷+答案解析·结构化Markdown·附RAG切分库

Internal agent platform on Amazon Bedrock AgentCore: interactive Claude Code cloud workspaces + headless agent kernels behind one portal, with publishing, scheduling, channels, memory, evals and governance

Engagement fact graph for autonomous pentest agents — typed facts, flag ledger, arena adapter, offline KB. Offensive fork of claude-fact-graph.

Material Design 3 and M3 Expressive packaged as an agent skill. Exact design tokens, per-component dp geometry, motion springs, and 2,587 reference images for Claude Code, Codex, Cursor, and Copilot.

Meme search & auto meme placement for AI agents. Remote MCP server, Agent Skill, CLI and REST API that let Claude Code, Cursor and Codex find reaction memes, GIFs and Chinese 表情包 by feeling and insert them into articles. 表情包搜索 · AI 配图 · 斗图 MCP

Community guides, megathreads, and resources for Hermes Agent by Nous Research. Mirrored from r/hermesagent.

AI Agent／大模型应用工程师中文学习与面试手册：10 章基础、18 章进阶、110 道详解题，配图解、追问与案例、Python 实验和复习工具。支持在线阅读与章节、题解双向跳转，用工具调用、RAG、评测和故障恢复案例练习工程判断。

一份只收 开源、可安装的视频类 Agent Skill 的精选索引——主打 Claude Opus 5.5 真正擅长的出片方式。

agent-flow: a live tree of the session's subagents beside the transcript, as a Claude Code mod (/flow)

Codex skill for auditing and cleaning local Claude/CC Switch/coding-agent environment residue

Five SKILL.md files that give AI agents UX expertise: evaluate interfaces, detect UX smells, audit AI interfaces, check flows. For Cursor, Windsurf, Claude Code and kx

Self-hosted remote control for Claude Code, Codex, OpenCode, Kimi Code, Grok Build and Cursor: run coding agents on your own machines and supervise them from anywhere, end-to-end encrypted.

Zhihu scraper and Markdown conversion utilities.

降低 LLM 角色扮演中软拒(soft refusal)与空回(empty response)的思路手册。核心：攻击外审模型注意力。纯文档项目，可在任意 Agent 框架复现。灵感来自类脑社区酒馆预设。

A curated collection of Claude Code agent skills that accelerate the entire vLLM development lifecycle.

A universal memory and knowledge layer for AI agents — structured enough for automation, readable enough for humans.

Install AI agent skills from Git into Claude, Codex, Cursor, and OpenClaw with a manifest, lockfile, and shared store

An Electron-based multi-project AI cockpit that orchestrates multiple Claude Code instances with a built-in AI supervisor for one-click dispatch, auto-monitoring, and report aggregation.

🔍 Model Context Protocol (MCP) tool for parsing websites using the Jina.ai Reader

Multi-Agent Deep Research System for Reddit

Fluent GPT App, an open-source, multi-platform desktop application that brings the power of GPT models to your fingertips. Designed with a sleek Fluent interface, it offers a unique and customizable chat experience on Windows, macOS, and Linux.

The Molt Autonomous Network (MAN) is an onchain agent swarm, coordinating to upgrade resources and achieve shared goals.

code for DyPS: Dynamic Parameter Sharing in Multi-Agent Reinforcement Learning for Spatio-Temporal Resource Allocation

Code for the NeurIPS 2023 Paper: Robust Multi-Agent Reinforcement Learning via Adversarial Regularization: Theoretical Foundation and Stable Algorithms

Large model based multi-agent platform for event reconstruction.

🔧 Base64 Proxy Tool - Cloudflare Workers代理工具，支持递归Base64解码、自定义UA和多层编码处理 A Cloudflare Workers-based proxy tool with recursive Base64 decoding, custom User-Agent, and multi-layer encoding support.

Solana Agent Example App

An open-source AI agent for infrastructure debugging.

Codebase context compiler for AI agents. Graph-ranked, token-budgeted, tier-compressed output. Smarter than the existing ones.

Formula WorkPaper and XLSX recalculation runtime for Node.js services and agent tools: edit cells, recalc, verify readback, persist JSON.

A lightweight nanobot/OpenClaw variant in pure Julia

Multiagent deep reinforcement learning research project

Common Chronicle turns messy context into structured, sourced timelines.

Mili is an artificial assistant that can make your life easier. It can perform a variety of tasks, such as navigating to a place, playing songs, and much more.

SPLAA is an AI assistant framework that utilizes voice recognition, text-to-speech, and tool-calling capabilities to provide a conversational and interactive experience. It uses LLMs available through Ollama and has capabilities for extending functionalities through a modular tool system.

Self-trained Large Language Models based on Meta LLaMa

XMPP Bot is a tiny little bot making the link between XMPP conversations and webhooks.

🚀 一键部署！支持语音聊天与数学逻辑处理! 基于各项LLM的 QQ 聊天机器人！支持 任意GPTLike的API，人设自定义,对话自然 🚀

DrFAQ is a plug-and-play question answering NLP chatbot that can be generally applied to any organisation's text corpora.

DEPRECATED: Use https://github.com/slack-php/slack-php-block-kit instead

Private AI Hub (P8Hub) - Host and use your own AI Services. Keep everything simple and private.

official repo for AAAI ALOHA chatbot

Knick is an open-source Python AI assistant 🧠

一个拥有长期记忆, 表情动作, 语音对话/打断/声纹识别, FunctionCall, 多模型支持的AI Waifu客户端.

Intelligent ChatBot built with Microsoft's DialoGPT transformer to make conversations with human users!

a stateless chat bot to perform natural language queries against the App Store top charts

💁 Example code for a blog post series about using a RAG system on a local codebase.

A proof-of-concept for a RAG to query the scikit-learn documentation

Open source NLA. Hosted service in private beta, signup now! Alternative to Zapier NLA

rudradb-opin-examples is for example implementations of the pip install rudradb-opin

Use Gemini Pro LLM via VertexAI to create an engaging quiz game incorporating TMDB API data

网络安全 LLM 智能体应用教程

A robust framework for building AI agents with multi-provider LLM support

Swarms for Javascript, the future of AI, where we leverage the power of autonomous agents to create 'swarms' of Language Models (LLM) that work together, creating a dynamic and interactive AI system.

Jarvis - Autonomous desktop agent

ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents

An exploration of using multiple agents collaborating to perform research

Atari gauntlet for RL agents

Implement Google Deep Minds DQN for multiple agents for a grid world environment where vehicles must pick up customers.

Run OpenClaw and OpenCode in the browser. Self-hosted, multi-user, Helm-deployable.

Multi-agent Task Allocation (WAFR2022)

Conflict-based search for optimal multi-agent pathfinding

CoMAL: Collaboration Collaborative Multi-Agent Large Language Models for Mixed-Autonomy Traffic

Prioritized Multi-Agent Path Finding using Safe Interval Path Planning (SIPP) and modifications

Multi-Agent framework for Godot

OpenReviewer is an automated multi-agent system for deep, evidence-driven research paper review, validation, and integrity auditing across methodology, experiments, citations, and cross-section consistency.

A native macOS dynamic island interface for the Pi coding agent. Real-time session monitoring, ambient notifications, and seamless multi-session management - all from your notch.

Claude Code power-user setup: 5 agents, 32 skills, notification hooks, workflow rules