A sophisticated workflow controller for managing continuous, multi-step AI subagent task lines without manual user re-prompting.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install subagent-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 subagent-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 Subagent Orchestrator?
The Subagent Orchestrator is designed to transform fragmented AI interactions into a seamless, continuous autopilot experience. Instead of stopping after a single run, this skill empowers the main controller to treat tasks as evolving lines of work, automatically re-dispatching subagents for UI, code, validation, and research roles. It is a vital addition to the library of Openclaw Skills, ensuring that complex project lifecycles move from inception to completion with minimal friction.
By leveraging a centralized task board and a strict return protocol, the orchestrator maintains context across different sessions. It balances the need for durable memory with the efficiency of ephemeral workspaces, cleaning up finished sessions while preserving essential metadata and stable collaboration rules. This allows developers to focus on high-level goals while the orchestrator handles the granular handoffs between specialized roles.
Subagent Orchestrator Use Cases
- Managing multi-stage UI and frontend integration pipelines that require sequential research and coding.
- Automating code refactoring workflows where logic changes must be immediately followed by validation subagents.
- Orchestrating long-running reconnaissance or documentation tasks that involve multiple data-gathering rounds.
- Coordinating cross-functional agent teams where one role's output serves as the direct dependency for the next.
How Subagent Orchestrator Works
- The main controller initializes by reading the active task line definitions from the AGENT_TASK_BOARD.md file.
- It identifies the next logical slice of work and spawns the specific subagent role required for that node.
- The subagent performs its task and returns a structured report containing status tags like autopilot, handoff, or blocked.
- Upon receiving the return, the controller updates the task board with the current progress, completed items, and risks.
- If the task line is set to autopilot, the controller advances the node and immediately dispatches the next role in the same turn.
- Completed session data is summarized into durable memory files before the temporary session environment is cleaned.
Subagent Orchestrator Setup
To integrate this skill into your environment, ensure your project includes the standard task board and protocol references. Install the skill using the following command:
openclaw install subagent-orchestrator
Next, initialize your project structure by creating an AGENT_TASK_BOARD.md and a references/SUBAGENT_RECORD_PROTOCOL.md to define your task nodes and communication standards. This setup ensures compatibility with other Openclaw Skills in your workflow.
Subagent Orchestrator Data Schema & Taxonomy
The orchestrator relies on a structured data schema to track state and ensure cross-role consistency. The primary data points include:
| Component | Location | Description |
|---|---|---|
| Task Board | AGENT_TASK_BOARD.md |
Tracks goals, current nodes, completed items, and next slices. |
| Session Memory | memory/YYYY-MM-DD.md |
Stores durable records of task goals, changed files, and result status. |
| Return Protocol | SUBAGENT_RECORD_PROTOCOL.md |
Defines mandatory fields like tag, line, node, and goal_status. |
| Artifact Management | Local Workspace | Temporary files and screenshots are deleted post-verification to maintain a clean environment. |
Subagent Orchestrator Advanced Features
- Anti-drop guards that verify every autopilot return against the task board to prevent workflow stalling.
- Mandatory delivery-proof rules requiring message IDs or channel links for tasks involving external platforms like Discord or Telegram.
- Automated cross-role dependency routing that links research, code, and validation roles without user intervention.
- Dynamic session cleanup that preserves durable memory records while purging intermediate artifacts.
- Promotion of stable collaboration patterns into the global MEMORY.md for long-term project intelligence.
SKILL.md
Loading
Related Openclaw Skills
Bailian TTS
A high-performance text-to-speech skill leveraging Aliyun Bailian for natural AI voice generation via command-line tools.
Model Usage (Linux)
A Linux-native tool to analyze AI model token consumption and financial costs from session logs.
Habit AI Health & Nutrition Tracker
A comprehensive health and nutrition tracking skill that leverages the Habit AI API to log meals, water, steps, and mental wellness.
Solana Development Skill
An opinionated end-to-end playbook for building, testing, and deploying high-performance Solana programs and React-based decentralized applications.
AllStock Data Query
A versatile stock market data retrieval tool supporting China A-shares, Hong Kong, and US markets via lightweight APIs and advanced SDKs.
Feishu Document Operations (User Access)
A comprehensive tool for interacting with Feishu documents using user-level authentication for reading, writing, and block management.
Featured*
