A structural framework for AI agents that separates identity, operational logic, and knowledge management for better scalability and consistency.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-os-three-layer
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-three-layer 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 comprehensive template designed for developers building sophisticated AI agents. It utilizes a three-layer architecture—Identity, Operations, and Knowledge—to decouple the agent's persona from its execution logic and data storage. This modular approach allows for more maintainable and robust implementations when creating complex workflows using Openclaw Skills.
By providing a standardized directory structure and configuration files, Agent OS ensures that every AI agent has a clear soul, a defined set of operational protocols, and a mechanism for long-term memory. This is particularly useful for developers who need to transition from simple prompts to full-featured autonomous systems that maintain context over time.
Agent OS Use Cases
- Creating custom AI assistants with persistent personalities and specific user profiles.
- Developing multi-agent systems where roles like CEO or specialized agents need clear operational boundaries.
- Implementing long-term memory and cross-session context for enterprise Openclaw Skills.
- Standardizing the development lifecycle of AI systems to ensure consistency across different projects.
How Agent OS Works
- The system initializes by setting up the three-layer directory structure: identity, operations, and knowledge.
- The Identity layer defines the core soul and persona of the agent, ensuring the AI understands its role and the user's requirements.
- The Operations layer manages the execution flow, including agent team definitions and system heartbeats to monitor health.
- The Knowledge layer acts as the agent's memory, storing shared context and long-term data to improve future interactions.
- Startup and verification scripts ensure the architecture is correctly configured and ready for deployment within Openclaw Skills.
Agent OS Setup
To get started with Agent OS, follow these steps in your terminal:
# Navigate to your agent-os directory
cd ~/agent-os
# Execute the startup script to initialize the system
./scripts/startup.sh
# Verify the architecture and file integrity
./scripts/verify.sh
Agent OS Data Schema & Taxonomy
The architecture is organized into specific directories to maintain the separation of concerns, which is a best practice for Openclaw Skills development:
| Layer | Key Files | Description |
|---|---|---|
| Identity | SOUL.md, IDENTITY.md | Defines personality, core values, and role configuration. |
| Operations | AGENTS.md, HEARTBEAT.md | Defines agent teams, operational instructions, and system health checks. |
| Knowledge | MEMORY.md, shared-context/ | Stores long-term memory and maintains state across different chat sessions. |
Agent OS Advanced Features
- Modular Identity-Operations-Knowledge (IOK) framework separation.
- Built-in heartbeat monitoring for maintaining operational stability in autonomous agents.
- Support for shared-context management to handle complex, multi-step workflows in Openclaw Skills.
- Role-specific instruction sets (e.g., ROLE-CEO.md) for orchestrating multiple sub-agents.
- Automated verification scripts to prevent architectural drift during development.
SKILL.md
Loading
Related Openclaw Skills
News Aggregator
An automated news intelligence agent that gathers, filters, and summarizes global technology and military news from authoritative sources.
Beetrade CLI
A professional toolset for managing Beetrade trading bots, strategies, and portfolio data via the beecli terminal interface.
Jable Video Downloader Lite
A lightweight tool for downloading Jable.tv content with automated actress-based organization.
Xiaohongshu Viral Copy Generator
An AI-powered content creation tool that generates structured, viral-ready copy for Xiaohongshu using proven engagement frameworks.
Best Practices
Enforce modern web development standards, security protocols, and code quality patterns within your AI-driven development environment.
MemOS Memory Recall
A persistent memory skill that enables AI agents to query and recall past interactions and user preferences from MemOS Cloud.
Featured*
