Build your own Cowork, AI Scientist and other SoTA Agents just by editing config files. Support anthropic skills. An infinite-horizon agent framework designed for long-running, complex tasks.

Long-running unsupervised agent framework

A lightweight Claude Code framework to set up multi-agent workflows, common repetitive tasks, and manage long-running projects

🔧 LRA - AI Agent长任务超稳定执行框架 ⚡ 轻量 | pip一键安装｜突破上下文限制 📌 7种状态管理 | Long-Running AI Agent Framework • One-click pip install｜Break context barriers • 7-status management

Long-running autonomous agent harness for Claude - 24/7 coding agent framework

一个让 AI 智能体能够长时间运行并完成复杂任务的编排框架 | An orchestration framework for AI agents to run for extended periods

LLM framework for long running agents

A framework for executing long-running AI agent tasks with structured feature lists and progress tracking.

AI Agent Framework for Long-Running Automatic Development

AI Agent Framework for Multi-Context Window Development

Long-running AI agent framework

Multi-agent harness framework for long-running application development with Claude Code

FRACTAL (Fractal, Recursive, Agentic, Context-aware, Task-driven, Autonomous, Layered) is a hierarchical framework for orchestrating teams of AI agents on complex software development tasks. It addresses context drift, serialization of parallel work, and cost inefficiency in long-running agentic sessions.

A Prompting Framework for Long-Running GitHub Copilot Agents

CLI framework for Claude Code - multi-session long-running agents

A public five-layer framework for long-running agent collaboration.

Framework for running and monitoring long-running Claude Code agents with self-healing capabilities

A verification framework for long-running agent tasks with milestone-based cumulative scoring

OpenClaw long-running agent framework - enables multi-session complex tasks with progress tracking

A framework for building agents with a focus on longer running, creative, reasoning tasks.

Convention-based harness framework for long-running Claude Code agents. Based on Anthropic's engineering blog.

Autonomous, long-running agent framework — pluggable LLMs (Claude, Gemini, Ollama), multi-channel (Telegram, WhatsApp, CLI), self-growing skills, pluggable memory backends

Demonstrates production-ready patterns for agentic orchestration using Microsoft Agent Framework. Covers structured output, tool-enabled agents, agent composition, reflective workflows, and agent handoff, focusing on state, delegation, observability, and long-running execution in real systems.

Memora is an AI Agent framework built around persistent file-based memory and structured task planning, designed for long-running, multi-step, real-world execution.

EnhanceAI: A production-grade framework for persistent, resumable AI agents. Replaces linear loops with an Execution Tree architecture and Hierarchical Memory (Episodic + Semantic). Engineered for deterministic task recovery, long-running workflows, and full-path auditability.