conductor

Orchestrate parallel task execution and sequential reviews in reactive scheduling loops.

1|Updated Jan 25, 2026
One-click install
npx skills add https://github.com/Zxela/claude-create --skill conductor-zxela
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: conductor
Source: https://github.com/Zxela/claude-create/tree/main/skills/conductor
Command: npx skills add https://github.com/Zxela/claude-create --skill conductor-zxela

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

The conductor skill orchestrates Phase 3 (Implementation) of the Homerun workflow by coordinating parallel implementer agents, sequential reviews, and failure handling to ensure efficient, reliable feature realization.

Core Features & Use Cases

  • Poll running tasks to detect completions, then process reviews sequentially
  • Spawn implementers in parallel within per-model and global concurrency limits
  • Handle failures by severity (retry, escalate, or block as needed) and refresh context periodically
  • Manage state and transitions between phases, preparing for Phase 4 when all tasks complete

Quick Start

Initialize and run the conductor within your workflow orchestration; monitor state.json to observe task progression and ensure pre-spawn checks are satisfied.

Frequently Asked Questions about conductor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate parallel task execution with dependency management?

Parallel task execution is orchestrated through a reactive scheduling loop that manages task dependencies, spawns implementers concurrently within per-model and global concurrency limits, and polls for completions.

How does failure handling work for parallel automation workflows?

Failure handling processes errors by severity, allowing the system to automatically retry, escalate, or block tasks within the orchestration loop to ensure reliable workflow progression.

What is the best way to manage state and context across parallel development tasks?

State and context are managed by periodically refreshing context during execution, validating traceability, and persisting task progression in a stateful format like state.json for monitoring.

Can I run sequential reviews while tasks are still executing in parallel?

Yes, the orchestration loop polls running tasks to detect completions and immediately processes their reviews sequentially, ensuring staged reviews happen alongside ongoing parallel implementation.

Do I need pre-extracted context to spawn parallel implementation agents?

Yes, task spawning requires pre-extracted context and pre-spawn checks to be satisfied, ensuring each parallel implementer agent has the necessary isolated data before execution begins.

When should I not use a parallel orchestration loop for task management?

You should avoid parallel orchestration for single-task workflows or simple scripts that lack dependencies, staged reviews, or the need for per-model concurrency controls and periodic context refresh.