Custom agents

Proof-driven Claude/Codex agent loop with paired execution, persistent resume, and adversarial security/performance gates.

An open-source AI agent for the browser. Bring any model. It plans in the side panel, acts on real pages through tools, and verifies every claimed result against the live DOM — with guardrails and reusable skills.

Claude Code best practices applied to application design. Interactive HLD/LLD visualizations, a DB-governed implementation example, and the same primitives as a runnable agent: the Governed Agent. LLM-agnostic, DB-governed, GDPR-ready.

Turn Android Termux into an autonomous AI agent and security orchestrator. Powered by a local ReAct framework with 50 system tools and MCP support, a background task │ scheduler, on-device ADB controls, and a unified Web UI & Telegram Bot.

You've probably been using Claude irresponsibly. Today, you can turn a new leaf.

Security control plane for AI agents — identity and delegation, capability policy, data-flow taint and a live audit trail, enforced over MCP. Guards a real Claude Code end to end.

LLM/ML 스터디·지식공유 큐레이션 — 검증된 강의·논문·자료 링크 (by @melocream · marblo.app)

Strata Studio is an agentic DAW that your AI agent (Claude Code, Cursor, Codex) can natively control so you can focus purely on making music.

[Corca / ML] Automatically solved Gandalf AI with LLM

基于 Mcp 架构的 ChatBI，是一种数据分析智能体的解决方案

根据文本相似度实现问答的聊天机器人（简单版）

Autoresearch with PhD-level workflows and modular agent skills. Built for the autonomous AI Scientist.

This repository is the official implementation of Learning Multi-Agent Coordination for Enhancing Target Coverage in Directional Sensor Networks.

🛒 E-commerce & Retail skill suite derived from iannuttall/claude-agents.

The Qude Framework open-source code allows users to create AI agents directly in their CLI, with Qude acting as a co-pilot to help build them. The Qude API enables interaction with the AI agents created using the Qude Framework's open-source code.

86 agent-executable skill packs converted from RefoundAI’s Lenny skills (unofficial). Works with Codex + Claude Code.

Gym environment for cooperative multi-agent reinforcement learning in heterogeneous robot teams

AI/LLM tool use (function calling) JSON Schema generator - a Kotlin multiplatform library

Demo application showcasing how to build and secure MCP servers and clients with Pomerium using contextual access policies.

Nexus - The Agentic Framework created by Talus

MCP server that allow LLM in agent mode builds n8n workflows for you

PIC: Permutation Invariant Critic for Multi-Agent Deep Reinforcement Learning

A resource to create a multi domain Dialog Act Tagger for conversational agents using publicly available data

Show ProgressBar on every view by calling an extension function

An Agent-Based Modelling (ABM) Framework for Unity3D

Your AI Copilot in Rust

Your AI command-line copilot

Squrve is a lightweight yet powerful framework for translating natural language into SQL over complex databases.

Official Code for "A Set of Control Points Conditioned Pedestrian Trajectory Prediction (AAAI 2023)"

Claude Code, natively inside Unity Editor. A dockable chat window with full project awareness, 60+ MCP tools, and zero alt-tabbing.

A reusable framework of agent skills that bring structure and repeatability to AI-assisted software delivery.

Embodied and organized multi-LLM-agent teams supporting communication for >3 agents. Source codes for the paper "Embodied LLM Agents Learn to Cooperate in Organized Teams".

ModelCraft is an agentic application designed for MacOS. It aims to make it easier to use LLM to complete task locally.

Codes accompanying the paper "DOP: Off-Policy Multi-Agent Decomposed Policy Gradients" (ICLR 2021, https://arxiv.org/abs/2007.12322)

Quickly deploy React code from Claude Artifacts to the internet.

A Reinforcement Learning (RL) simulation environment built for training and evaluating defensive and offensive autonomous models on simulated and emulated networks, with a focus on modularity and configurability.

A full-stack AI chat application powered by Amazon Bedrock and Strands Agents

npx CLI that scaffolds blockchain-registered AI agents with on-chain identity (ERC-8004), A2A communication, MCP tool serving, and optional USDC micropayments across EVM and Solana.

🧩 Tibo – CLI tool for codebase indexing, embedding and natural language querying.

Multi Agent Traffic Scenario Gym: A scenario-based training and evaluation framework for CARLA.

Source codes for the paper "COMBO: Compositional World Models for Embodied Multi-Agent Cooperation"

MARS is shortened for Multi-Agent Research Studio, a library for mulit-agent reinforcement learning research.

The Karpathy treatment for OpenClaw - stripped to ε. 515 lines of Python, 6 files, one dependency.

LLM driven natural language to generate summaries, bibliography lists, download PDF reports autonomously and extract and visualise open domain data (curently uses GPlates, Macrostrat and Mindat).

layer that lets you connect any agent, any tool, any api together.

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.

linked of Romanian fiscal authority(MFP-ANAF-GOV-RO)

OpenClaw skill: AI agent takes over Android/Linux devices via ADB+SSH — see screen, send inputs, post content, play games autonomously

Code and data for the paper: On the Resilience of LLM-Based Multi-Agent Collaboration with Faulty Agents

Agent Skill to help convert transformer LLMs to mlx-lm

Used OpenClaw to set up an autonomous multi-agent system that reads OpenClaw's own source repo, gathers info from the web, and writes a full book about the platform.

Semantic memory system with knowledge graph, spreading activation, embedding-based recall, autonomous dream consolidation, and C++ LSTM+kNN pattern learning for any /MCP and the Hermes Agent.

[NAACL 2025] AgentMove: A Large Language Model based Agentic Framework for Zero-shot Next Location Prediction.

MCP server for spreadsheet analysis and editing. Slim, token-efficient tool surface designed for LLM agents.

Local Retrieval-Augmented Generation (RAG) plugin for Claude Code that combines Chroma db vector embeddings with intelligent info retrieval with Multi-Agent Framework (MAF) orchestration for context-aware development assistance. Uses Open Source / Free frameworks. Implements bridge to Claude Code CLI so no token use. And it's easy to setup.

Motor de Inteligencia de Enjambre para el mercado hispano 🐟 | Sube cualquier informe, predice el futuro al instante | Adaptación en español por EnPíxeles Lab

Standalone web app that visualizes Claude Code sessions as pixel art characters in a virtual office. Fork of pablodelucca/pixel-agents.

A curated list of 260+ papers and resources on Agentic Security. Based on the survey covering the transition from passive LLMs to autonomous agents, structured around three pillars: Applications (Red/Blue Teaming), Threats (Injection, Poisoning), and Defenses (Guardrails, Verification).

A to-do list that runs your Claude Code agents — capture anywhere, dispatch in parallel, review from your phone.

Obsidian plugin to visualize agent skill structures (OpenClaw / Claude Code) in graph view

A CLI for DaVinci Resolve 18+ — project / media / render / timeline control for humans and AI agents.

CRISP — The missing layer between vibe coding and building something people actually want.

Autonomous knowledge base plugin for Claude Code - captures reserch, ideas, and decisions into an interlinked wiki with reserch-on-miss, semantic search, and a Wikipedia-style web UI. Knowledge compounds as you work.

MCP server that exposes a V8 JavaScript runtime as a tool for AI agents like Claude and Cursor. Supports persistent heap snapshots via S3 or local filesystem, and is ready for integration with modern AI development environments.

Supercharge your GitHub Copilot with multiple AI providers

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.

A toolkit for frontier forecasting.

A behavioral governance framework for GitHub Copilot agents — skills.md and instructions to make AI coding agents think like senior engineers.

Command-line tool for the Chrome DevTools Protocol (CDP): debug targets, console, traces, heap snapshots, and JavaScript CPU profiling for Chrome, Node.js, and React Native.

Recallium is a local, self-hosted universal AI memory system providing a persistent knowledge layer for developer tools (Copilot, Cursor, Claude Desktop). It eliminates "AI amnesia" by automatically capturing, clustering, and surfacing decisions and patterns across all projects. It uses the MCP for universal compatibility and ensures privacy

OpenHermit is the open-source platform for deploying fleets of AI agents as production services — durable state, sandboxed execution, managed at scale, and the channels you already use.

Open-Ended Long-Horizon Text Game Generation for Test-Time Continual Learning Agents

Agent Skills for traceable requirements, independent verification, human approval gates, and auditable AI-assisted engineering. Supports Claude Code, Cursor, VS Code, and GitHub Copilot; ISO 9001/27001 aligned (design phase), GxP-aware.

Local-first MCP server that gives coding agents structured context packets, code/schema facts, and diagnostics - backed by a local SQLite store.

Multi-tier firewall for AI agents — blocks prompt injections, jailbreaks, and scope violations. Local tiers first; LLM judge only when uncertain.

Spec-driven development and context engineering for Claude Code, Cursor, Codex, and GitHub Copilot — backed by project context in Git.

Toolkit for Seamlessly Enabling RL Training on Any Agent with Bedrock AgentCore.

Make AI-generated prose sound human. 29-pattern catalog + 8-rule operating system + 8-dimension scoring rubric. Works with Claude Code, OpenCode, Copilot CLI, Codex, Gemini, Goose and Hermes.

productivity-focused sandboxing for Linux

Device Context Protocol — bridge LLM agents to physical devices. Sub-50-byte frames, 27.6KB flash / 0.6KB RAM measured on ESP32, capability-scoped and safe by design. Complementary to MCP. Paper: arXiv:2605.26159

This is a darkweb forums tracker that monitors forum posts and sends alerts to Discord

ConDB: The KV-Cache Native Context Database

Hermes AI coding agent for VS Code — streams chat, runs tools, manages sessions, switches models (Claude/Codex), and tracks context live via ACP.

한국투자증권 OpenAPI Python 래퍼 — CLI · Python SDK · LLM Agent · 실시간 WebSocket. 문서: unohee.github.io/kis-agent

44 open-source financial skills for AI agents — P&L, budgeting, tax prep, debt payoff, and more. Works with Claude Code, Wilson, Cursor, Codex, and Paperclip.

EvE is an open-source framework for co-evolving ensembles of coding agents. Wrap Codex into an evolutionary loop that co-evolves solvers and agent guidance.

Create and install AI agents to automate your tasks

Your last agent.

Code repo for the paper: Attacking Vision-Language Computer Agents via Pop-ups

🧿 MythOS 2026: Claude-Powered AI Narrative Engine & Sub-Agent Framework for Epic Storytelling

Agent-as-a-Judge grading framework for evaluating AI outputs/deliverables

Elevate your AI assistants (like Claude & Cursor) into autonomous cybersecurity experts. Pentester-MCP integrates 200+ pentesting tools via the Model Context Protocol (MCP) using a secure Docker sandbox.

AI-powered QA testing framework that uses LLMs (Claude or GPT) to test web apps, CLI tools, and TUI programs from markdown story cards, returning structured pass/fail verdicts with evidence.

Multi-agent debate system for Hermes Agent — spawn custom-composed expert agents to debate any question in structured rounds, producing visible transcripts and synthesized recommendations. Pure SKILL.md, zero new infrastructure.

企业级5-Agent智能会议助手系统 | Pipeline+并行编排 | Python/Java/Go三语言实现 | 面试宝典(八股文+STAR法+简历模板)

Build the system that prompts your agents. A teaching repo for loop engineering: chapters, an annotated reading list, copy-paste prompts, a runnable example, and a portable agent skill

All-in-one AI 研究工作台：把 Codex / Claude Code / Cursor 变成有长期记忆的研究系统。wiki 摄入、研究闭环、快速筛选、假设追踪、播客、日报六大能力，零 API key 可跑基座。

TripStar multi-agent travel planner powered by Spring Boot 4, Spring AI Alibaba, StateGraph, multimodal note understanding and AMap