Agent Conductor 🎼

You conduct. Agents perform.

A skill for orchestrating coding sub-agents (Claude Code, Codex, Cursor, Gemini Code, and more) to maximize throughput on implementation tasks. Keep your orchestrating session lean — it plans, decides, and validates. Sub-agents do the execution.

Features

• 🤖 Agent-agnostic — works with Claude Code, Codex, Cursor, Gemini Code, or any CLI-based coding agent
• 📋 Ready-to-use dispatch template — copy, fill in, send
• 🔀 Parallel coordination patterns — data sharding, division of labor, hybrid pipelines
• ♻️ Checkpoint / resume — built-in pattern for batch tasks > 100 items
• ✅ Acceptance checklist — never trust "done" without verifying

Install

Option 1: market-cli (recommended)

npx @lobehub/market-cli skills install agent-conductor --agent open-claw
# or for Claude Code:
npx @lobehub/market-cli skills install agent-conductor --agent claude-code

Option 2: Manual

# OpenClaw
git clone https://github.com/AICodeLion/agent-conductor ~/.openclaw/workspace/skills/agent-conductor

# Claude Code
git clone https://github.com/AICodeLion/agent-conductor .claude/skills/agent-conductor

# Codex
git clone https://github.com/AICodeLion/agent-conductor .agents/skills/agent-conductor

Quick Start

1. Pick your agent — set your invoke command (e.g. claude, codex, cursor-agent)
2. Fill the dispatch template from SKILL.md
3. Choose execution mechanism based on estimated duration (< 5 min / 5–30 min / > 30 min)
4. Run the acceptance checklist after completion

When to Use

Dispatch when the task involves:

• Writing or modifying files (even one line)
• Running scripts or processing data
• Execution time > 10 seconds
• Batch operations over multiple items

Patterns

See references/patterns.md for:

• Data sharding across parallel agents
• Division of labor between agents
• Serial pipelines with dependencies
• Hybrid patterns (serial + parallel)
• Checkpoint/resume for long batch tasks
• Domain examples: data science, web dev, document processing

Supported Agents

Agent	Invoke Command
Claude Code	claude '<task>'
OpenAI Codex	codex '<task>'
Cursor Agent	cursor-agent '<task>'
Gemini Code	gemini-code '<task>'
Custom	your-agent-cmd '<task>'

License

MIT