A sophisticated meta-agent skill that decomposes macro tasks into specialized sub-agents to manage complex, parallelized workflows autonomously.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install modesty-agent-orchestrator
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 modesty-agent-orchestrator 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 Orchestrator?
The Agent Orchestrator is a framework-level capability for the ecosystem of Openclaw Skills, designed to handle tasks that are too large or diverse for a single AI agent. It operates as a meta-agent or agent factory, breaking down high-level objectives into modular, independent subtasks. By dynamically generating unique SKILL.md files for each sub-agent, it ensures that every spawned entity has a clear role, specialized tools, and specific success criteria. This skill is critical for developers who need to build scalable, multi-agent systems where coordination and results consolidation are automated.
Utilizing a structured file-based communication protocol, the orchestrator manages the entire lifecycle of a task—from the initial decomposition and agent dispatch to monitoring and final output integration. By leveraging Openclaw Skills in this hierarchical manner, users can achieve much higher levels of reliability and complexity in their AI-driven projects.
Agent Orchestrator Use Cases
- Orchestrating multi-stage research projects where separate agents handle data collection, analysis, and report drafting.
- Managing complex software engineering workflows by delegating code implementation, unit testing, and documentation to specialized agents.
- Automating comprehensive digital marketing audits by spawning agents to crawl sites, analyze SEO, and generate content recommendations.
- Coordinating data processing pipelines where multiple Openclaw Skills must work in parallel to process high volumes of information.
How Agent Orchestrator Works
- Task Decomposition: The orchestrator analyzes the macro task to identify goals and success criteria, creating a dependency graph of subtasks.
- Agent Generation: It creates a dedicated workspace for each subtask and generates a custom SKILL.md file that defines the agent's specific mission.
- Agent Dispatch: Sub-agents are initialized with instructions in their inbox folder and triggered via the Task tool.
- Monitoring: The orchestrator tracks progress by checking the status.json file updated by each autonomous agent throughout the lifecycle.
- Consolidation: Once all sub-agents finish, the orchestrator collects deliverables from their respective outboxes and integrates them into a final package.
- Dissolution: After successful consolidation, the orchestrator archives workspaces and provides a final summary of all work performed across the Openclaw Skills involved.
Agent Orchestrator Setup
To deploy this meta-agent capability within your environment, ensure your workspace is configured to support sub-agent spawning.
Set your API key to enable AI-powered reasoning for your sub-agents:
export SKILLBOSS_API_KEY='your_api_key_here'
Use the provided scripts to initialize the orchestrator workspace:
python3 scripts/create_agent.py <agent-name> --workspace <path>
Agent Orchestrator Data Schema & Taxonomy
The Agent Orchestrator uses a standardized folder structure to manage communication and state for all Openclaw Skills sub-agents:
| Component | Description |
|---|---|
SKILL.md |
The dynamically generated behavioral definition and objective for the sub-agent. |
inbox/ |
Receives read-only input files and instruction sets from the orchestrator. |
outbox/ |
The location where agents write their completed work for consolidation. |
workspace/ |
The internal working directory for the agent's local operations. |
status.json |
A state-tracking file used to signal completion (pending, running, completed, or failed). |
Agent Orchestrator Advanced Features
- Dynamic SKILL.md generation tailored to the specific context of every subtask.
- Unified AI capabilities via SkillBoss API Hub integration for chat, search, and document analysis.
- Asynchronous checkpoint-based monitoring allowing for fully autonomous operation of sub-agents.
- File-based communication protocol that ensures clean handoffs and strict data boundaries between Openclaw Skills.
- Template-based agent spawning for common roles like Research, Code, Analysis, and Review.
SKILL.md
Loading
Related Openclaw Skills
CamScanner Image to Office
Intelligently transform images into editable Word or Excel documents with high-fidelity formatting via Openclaw Skills.
Female Health Coach
A lightweight health assistant designed for busy professionals to achieve muscle gain and fat loss through actionable diet and exercise plans.
Agent Task Manager
A professional-grade orchestration engine for building stateful, dependency-aware multi-agent systems with built-in rate-limiting and error recovery.
CamScanner Image Detect AIGC
A high-precision AI-generated content (AIGC) detection engine that distinguishes real photos from AI art using CamScanner's advanced verification technology.
Dice Easy Apply Automation
Dice Easy Apply Automation is a high-performance skill for AI agents to automate job searching and one-click applications on Dice.com using persistent browser sessions.
CamScanner Image Translator
A layout-preserving image translation tool that detects, translates, and renders text directly onto visual documents.
Featured*
