subagent-orchestrator

What This Skill Does

The subagent-orchestrator is the central engine for OpenClaw's autonomous workflows. It enables the AI to move beyond single-turn interactions by managing multi-step, continuous autopilot task lines. Instead of waiting for human intervention between every task, this skill empowers the agent to autonomously spawn, monitor, and chain subagents (such as coders, validators, or scouts) to drive complex projects to completion. It maintains the state of the task board, handles role-based routing, and ensures that intermediate data is archived while keeping the workspace clean of unnecessary artifacts.

Installation

To integrate this skill into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/hackjsw/subagent-orchestrator

Use Cases

• Continuous Feature Development: Automate the handoff from technical research to code implementation and finally to UI verification without constant user prompts.
• Refactoring Pipelines: Orchestrate a sequence where a 'scout' analyzes code debt, a 'coder' performs the refactoring, and a 'validator' runs tests.
• Automated Documentation Updates: Sync project code changes with documentation updates by triggering documentation-focused subagents immediately after code commits.
• Complex Bug Resolution: Automatically route from bug discovery to root cause analysis, patch application, and verification.

Example Prompts

1. "Initialize an autopilot line to build the authentication module; use the orchestrator to manage the research-to-code flow."
2. "Update the AGENT_TASK_BOARD to reflect that the UI phase is complete and trigger the validation subagent."
3. "Start an autopilot session to refactor the landing page; route tasks through the appropriate roles and keep the project memory updated."

Tips & Limitations

• Maintain the Board: Always ensure AGENT_TASK_BOARD.md is updated. It is the single source of truth for the orchestrator.
• Be Concise: Force your subagents to use the short-report protocol. Excessive detail in reports slows down the orchestration flow.
• Memory Hygiene: Remember that this skill cleans up sessions. If you need to keep long-term context, ensure your subagents are configured to write to the memory/ directory before the orchestrator closes the session.
• Safety: Because this skill can trigger multiple agents, ensure that your environment has sufficient resource limits, as infinite loops of automated subagent creation can consume significant system resources if not properly constrained by task-board limits.