A powerful orchestration framework to manage and delegate implementation tasks to multiple AI coding sub-agents simultaneously.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-conductor
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-conductor 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 Conductor?
Agent Conductor is a high-level orchestration skill designed to separate planning from execution in AI-driven development. By leveraging this tool within Openclaw Skills, developers can maintain a lean orchestrator session that focuses on high-level architecture and decision-making while routing intensive implementation work—such as file modifications, script execution, and data processing—to specialized coding sub-agents like Claude Code or Gemini. This separation of concerns ensures higher throughput and prevents the primary session from becoming bogged down by long-running execution tasks.
The system is entirely agent-agnostic, allowing you to plug in any CLI-based coding tool using simple dispatch templates. Whether you are managing complex multi-stage workflows or parallelizing batch operations, this skill provides the necessary structure to verify outputs and handle errors reliably. By utilizing Openclaw Skills for orchestration, you ensure that the primary agent conducts while the sub-agents perform the heavy lifting.
Agent Conductor Use Cases
- Writing or modifying multiple code files across a repository.
- Executing long-running data processing pipelines or scripts.
- Batch processing large datasets where parallel execution is required.
- Decomposing complex multi-stage projects into verifiable implementation steps.
How Agent Conductor Works
- The orchestrator identifies a task requiring execution, such as modifying a file or running a script.
- A dispatch template is populated with context, requirements, and specific acceptance criteria.
- The task is routed to a sub-agent (e.g., Claude Code) via a CLI command.
- The sub-agent executes the task in the foreground or background based on the estimated duration.
- Once the sub-agent signals completion, the orchestrator runs a verification checklist to ensure data integrity and success.
Agent Conductor Setup
Define your preferred sub-agent command using the AGENT_CMD environment variable or direct configuration within your Openclaw Skills environment. Ensure your system has the necessary CLI tools installed.
# Example: Setting up Claude Code as the primary sub-agent
export AGENT_CMD="claude"
# Example: Running a task via the orchestrator using a PTY
# exec pty:true command:"AGENT_CMD 'Create a React component in src/components/Button.tsx'"
Agent Conductor Data Schema & Taxonomy
| Component | Description |
|---|---|
| Dispatch Template | Markdown structure containing requirements, context, and criteria. |
| Execution Logs | Captures stdout/stderr from sub-agents for debugging. |
| Progress Tracking | Uses progress.json to manage state during multi-stage batches. |
| Acceptance Checklist | A validation manifest used to confirm output file existence and record counts. |
Agent Conductor Advanced Features
- Multi-agent support for parallelizing tasks across different AI models.
- Background execution modes with configurable timeouts for long-running scripts.
- Task decomposition logic to split large projects into independently verifiable stages.
- Checkpoint and resume capabilities using local state tracking for batch operations.
SKILL.md
Loading
Related Openclaw Skills
4chan Reader
A specialized data extraction tool for browsing 4chan boards and converting thread discussions into structured text files.
Reading & Knowledge Agent
A judgment-free knowledge companion that provides expert insights into space, the universe, and history while building your personal learning database.
Kindergarten Teaching Assistant
An AI-powered educator specializing in the British EYFS framework and Reggio Emilia pedagogy for children aged 45 days to 2 years.
Fitness Support
A specialized AI fitness coach for building muscle with dumbbells and tracking progressive overload.
agent-pack-n-go
A powerful migration tool that clones your entire OpenClaw Agent environment, including memory and credentials, to a new device via automated SSH scripts.
STM32CubeMX CLI
A professional automation skill for configuring STM32 microcontrollers and generating project code via a headless command-line interface.
Featured*
