Awesome GitHub RepositoriesAgent Execution Environments

This project is a structured educational resource and technical guide for designing and implementing autonomous systems using large language models. It provides a comprehensive curriculum and code samples focused on agentic design patterns, autonomous development, and the creation of systems capable of planning and executing multi-step tasks. The resource details the implementation of agentic retrieval-augmented generation, where models autonomously plan and refine data searches. It covers a wide array of orchestrators and design patterns, including metacognitive reflection for self-correctin

Nanoclaw is an LLM agent orchestrator and multi-platform chat gateway designed to deploy and manage isolated AI agents. It provides a containerized runtime that executes agents within sandboxed Linux containers, ensuring filesystem and state isolation through dedicated workspaces and host bind-mounts. The project distinguishes itself through a unified routing pipeline that connects agents to diverse messaging platforms, including WhatsApp, Discord, Slack, Telegram, Signal, and iMessage. It integrates the Model Context Protocol to extend agent capabilities via managed external data and functio

OpenClaude is an LLM orchestration interface and multi-provider AI gateway that connects various AI providers and local models to an integrated tool suite. It functions as an agentic tool execution environment and a system for AI-powered code editor integration, enabling in-editor chat and automated coding tasks. The project provides a gRPC AI agent service that exposes model capabilities and file editing tools to external applications as a headless service. It also includes a configuration layer for managing provider credentials and routing specific agents to different model APIs. The syste

This project is a Python framework for building autonomous, event-driven agent systems. It provides a unified runtime for orchestrating multi-agent workflows, managing persistent conversation state, and executing code within secure, isolated sandbox environments. The framework is designed to handle complex task delegation, allowing agents to invoke other agents as tools while maintaining context across multi-turn interactions. The framework distinguishes itself through its deep integration with the Model Context Protocol, enabling agents to connect to external data sources and remote services

AgentScope is a multi-agent framework and orchestration platform designed for building and coordinating teams of language model agents. It provides a system for managing multiple agents that collaborate to solve complex tasks through structured communication and state sharing. The project distinguishes itself with a focus on production-ready deployment and security, featuring a multi-tenant hosting service that ensures session isolation between different users. It includes a sandboxed tool execution environment and fine-grained permission controls to manage how agents access system resources

This project is a Docker educational resource and a collection of practical examples designed for learning containerization technologies. It serves as a guide for understanding container fundamentals, including the creation and management of custom images and the use of registries. The repository provides specialized references for container security hardening, such as managing kernel privileges and implementing supply chain security. It also includes tutorials for multi-container orchestration and a DevOps guide focused on CI/CD automation and image optimization. The material covers a broad

NemoClaw is an LLM agent orchestrator and sandboxed execution environment designed to deploy and manage the lifecycles of large language model agents. It provides a secure runtime that isolates persistent agents from the underlying host system to ensure operational security. The system includes a secure LLM inference gateway that acts as a managed routing layer, securing communication between AI agents and inference engines to prevent unauthorized access. It also integrates with NVIDIA OpenShell to run specialized agents within a secure shell environment. Operational control is provided thro

OpenFang is an operating system for LLM agents designed to orchestrate autonomous agents with built-in task scheduling, tool sandboxing, and multi-model routing. It provides a secure AI execution environment that integrates prompt injection scanning, cryptographic audit trails, and resource metering to ensure controlled processing. The platform distinguishes itself through a comprehensive security architecture, featuring fuel-metered tool sandboxing and an immutable activity audit trail based on cryptographic hash-chains. It implements high-assurance identity verification via signed manifests

Claude Quickstarts is a development framework and collection of reference implementations designed for building autonomous agents. It provides the foundational patterns necessary to orchestrate multi-agent workflows, enabling models to perform complex, multi-step tasks across software engineering, customer support, and computer-use domains. The platform distinguishes itself through specialized capabilities for desktop and browser automation, allowing agents to interact with graphical interfaces by capturing visual context and executing precise mouse and keyboard inputs. It includes robust inf

DeepCode is an agentic development framework designed to orchestrate autonomous AI agents for software engineering tasks. It functions as a multi-agent workflow orchestrator that translates natural language requirements into functional codebases by coordinating specialized agents for architectural planning, intent analysis, and implementation. The platform integrates multiple language models to power these automated routines, providing a unified environment for complex development projects. The system distinguishes itself through its ability to transform academic research papers into executab

E2B is a cloud-based infrastructure platform designed to provide secure, isolated execution environments for code and shell commands. It functions as an ephemeral orchestrator that provisions lightweight virtual machines, allowing developers and autonomous agents to run untrusted processes within a sandbox that is completely separated from the host system. The platform distinguishes itself through its focus on programmable, serverless workspaces that support the full lifecycle of cloud-based development. By utilizing hardware-level isolation and snapshot-based resumption, it enables the near-

OpenBrowser is an AI web agent toolkit and automation framework designed to translate natural language instructions into executable browser workflows. It functions as a headless browser controller and orchestrator, enabling the creation of autonomous agents that navigate websites, interact with elements, and extract data using plain English commands. The system features a sandboxed execution environment that utilizes domain whitelists and memory limits to ensure secure web interaction. It distinguishes itself through a command-line interface for triggering autonomous tasks with configurable m

This project is a comprehensive framework for the orchestration, evaluation, and context management of large language model agents. It provides a set of architectural patterns and standards for designing agent interactions, integrating external tools, and establishing memory architectures to persist knowledge across sessions. The system focuses on optimizing the limited memory of language models through token-aware context compression and filesystem-based context offloading. It incorporates secure execution environments using sandboxed virtual machines and isolated containers to safely run ba

Upsonic is a Python framework and orchestrator for building autonomous AI agents. It provides the infrastructure to develop self-operating systems that execute complex workflows and manage tasks independently using a scripting language. The project functions as an agentic tool integration layer, connecting agents to third-party data sources and external service APIs through standardized communication protocols. To ensure security, it includes an isolated execution environment that restricts shell and file operations to a specific workspace to prevent path traversal and dangerous commands. Ad

This project is an LLM coding agent orchestrator and AI software engineering platform designed to manage fleets of agents that autonomously solve issues, handle pull requests, and fix CI failures. It functions as an agentic CI/CD automator and parallel workflow manager, coordinating the end-to-end development lifecycle from initial ticket tracking to final code merging. The system is distinguished by its modular plugin framework and isolated worktree management, which allow multiple agents to work on separate coding tasks simultaneously without file system conflicts. It utilizes role-based mo

Yao is an LLM agent framework and low-code web app builder designed for orchestrating autonomous AI agents. It provides a platform to design, deploy, and coordinate agents with specialized personas that can plan tasks, utilize external tools, and execute multi-stage pipelines. The project distinguishes itself through a Model Context Protocol server for connecting assistants to external binaries and HTTP services, and a gRPC remote execution engine that allows agents to manage remote servers and devices. It includes a model-agnostic provider bridge that supports dynamic switching between vario

OpenShell est un framework de sécurité et un runtime d'exécution sandboxé pour les agents IA autonomes. Il fournit des environnements isolés utilisant des conteneurs et des machines virtuelles pour protéger l'infrastructure hôte et les données sensibles contre l'accès non autorisé pendant l'exécution de l'agent. Le système se distingue en combinant le passthrough accéléré par matériel pour l'accès GPU hôte avec une passerelle de sécurité qui intercepte les appels d'API du modèle. Cette passerelle gère les identifiants en supprimant les informations de l'appelant et en injectant des secrets backend, garantissant que les clés d'API sensibles restent hors du système de fichiers local. La plateforme couvre de larges domaines de capacités, incluant l'application de politiques déclaratives pour les restrictions de système de fichiers et de réseau, la gestion des identifiants d'agent et la surveillance de l'activité en temps réel. Elle prend en charge le provisionnement d'environnements via des images de conteneurs personnalisées, des répertoires locaux ou des catalogues communautaires.

Microsandbox est un runtime microVM sandbox et un exécuteur de code isolé matériellement conçu pour exécuter du code non fiable. Il fonctionne comme un gestionnaire de machine virtuelle embarqué qui permet aux applications de générer et de contrôler des machines virtuelles légères directement dans le code sans avoir besoin d'un démon en arrière-plan. Le système fournit un environnement d'exécution sécurisé pour les agents IA en exposant des contrôles serveur qui leur permettent d'exécuter des outils et de gérer des fichiers. Il utilise des formats d'image de conteneur standard et des workflows de volumes pour initialiser les machines virtuelles invitées et implémente un mécanisme de gestion des secrets qui empêche les clés sensibles d'entrer dans la mémoire de la machine virtuelle. La plateforme couvre le cycle de vie complet de l'orchestration de charges de travail isolées, incluant la création, la surveillance et la suppression des environnements. Elle inclut des capacités de surveillance des ressources CPU et mémoire hors bande, la mise en cache d'images invitées, et l'exécution de commandes immédiates ainsi que de sessions détachées en arrière-plan.

Hatchet is an open-source durable workflow engine and task orchestration platform. It provides a framework for building and executing fault-tolerant, multi-step pipelines as directed acyclic graphs (DAGs), with automatic retries, scheduling, and real-time observability. The system is built around durable task checkpointing, which persists execution state after each step so work can resume from the last checkpoint after a worker crash or restart, and it supports event-driven task resumption that pauses a task until a matching external event arrives. The platform distinguishes itself through it

Rivet est une infrastructure distribuée pour gérer le cycle de vie, l'adressage et la persistance d'acteurs stateful et de moteurs d'exécution durables. Il fournit un sandbox de processus distribué qui exécute la logique applicative au sein d'isolats légers, garantissant l'isolation des ressources et des démarrages à froid rapides. Le système est conçu pour coordonner des opérations multi-étapes en utilisant des files d'attente persistantes et des minuteurs pour garantir l'achèvement fiable des tâches dans des environnements distribués. La plateforme permet spécifiquement l'orchestration d'agents IA stateful qui maintiennent une mémoire et un état persistants à travers des interactions de longue durée et des workflows complexes. Elle se distingue par un framework de synchronisation d'état WebSocket qui lie les composants d'interface utilisateur frontend aux processus stateful distants via une communication bidirectionnelle en temps réel. Le système couvre un large éventail de capacités, incluant l'adressage hiérarchique d'acteurs, un runtime « hibernate-on-idle » pour l'optimisation des ressources, et une couche de persistance enfichable pour des backends de stockage modulaires. Il inclut également des outils pour le débogage de session active, la surveillance de l'état d'exécution en temps réel, et des options de déploiement automatisé pour l'edge, le cloud ou l'infrastructure privée. Le projet est implémenté en Rust et prend en charge le développement d'acteurs multi-langages.