A persistent operating system for OpenClaw agents that enables long-term memory, capability learning, and multi-agent project coordination.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-os
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install agent-os using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Agent OS?
Agent OS is a foundational layer designed for building sophisticated multi-agent systems. Unlike traditional stateless frameworks, this system provides a persistent execution environment where agents maintain their own history, learn from past successes, and coordinate on complex task sequences. By leveraging these Openclaw Skills, developers can create autonomous agents that survive restarts and maintain continuous state across long-running, multi-phase projects.
The system focuses on transforming high-level goals into actionable task sequences, ensuring that every agent involved knows exactly what to do based on its specific capabilities. This architectural approach reduces redundant API calls and ensures a more cost-effective and reliable agentic workflow.
Agent OS Use Cases
- Executing complex multi-phase projects such as simultaneous research, design, and development.
- Maintaining long-term agent memory to avoid redundant context resets and reduce token costs.
- Coordinating multiple specialized agents within a single project to prevent duplication of work.
- Resuming interrupted tasks mid-project seamlessly without losing execution progress.
How Agent OS Works
- Register specialized agents with specific capabilities such as research, design, or coding.
- Utilize the TaskRouter to decompose high-level project goals into a sequence of executable tasks.
- Match and route specific tasks to the best-fit agents based on their capability scores and history.
- Execute tasks through the scheduler while tracking progress and persisting state to the local file system.
- Automatically record lessons learned in the agent memory to improve future performance and success rates.
Agent OS Setup
Install the package via the CLI:
clawhub install nova/agent-os
Initialize the system in your project:
const { AgentOS } = require('agent-os');
const os = new AgentOS('my-project-id');
// Register your agents
os.registerAgent('dev', 'Developer Agent', ['development']);
os.initialize();
Agent OS Data Schema & Taxonomy
Agent OS organizes all persistent state within a dedicated data/ directory using a clear JSON-based taxonomy:
| File Pattern | Content Description |
|---|---|
[agent-id]-memory.json |
Stores agent-specific knowledge, lessons learned, and success rates. |
[agent-id]-state.json |
Tracks current agent status, active tasks, and identified blockers. |
[project-id]-project.json |
Contains the master project task list, dependencies, and overall progress. |
Agent OS Advanced Features
- Persistent agent memory for long-term capability improvement and context retention.
- Smart TaskRouter for automated goal decomposition and complex dependency tracking.
- State persistence ensures that all Openclaw Skills projects survive system restarts or crashes.
- Real-time execution tracking with a live progress board and detailed status reporting.
- Specialized agent registration allowing for highly granular capability-based task routing.
SKILL.md
Loading
Related Openclaw Skills
Shared Pinecone RAG
A centralized Retrieval-Augmented Generation (RAG) system that provides shared semantic memory and document querying across multiple AI agents.
Movie Producer Scene
A sophisticated prompt engineering tool that generates high-end cinematic scene packages, shot plans, and production-ready briefs in a professional Hollywood voice.
Hybrid DB Health
A diagnostic utility designed to validate the dual-layer database architecture used by AI agents for persistent memory and operational tasks.
Stoic Scope Creep
A practical framework applying Stoic philosophy to help developers and managers maintain effectiveness when project boundaries expand unexpectedly.
Build Transparency Dashboard
Automatically scaffold a public build dashboard that displays live commit data and project progress from a private GitHub repository.
Drift
Drift is a specialized conversation protocol that allows AI agents to maintain continuity by leaving questions and reflections for future instances of themselves.
Featured*
