How does pm-epic-start handle task dependencies?

The skill parses your epic documentation for 'depends_on' and 'parallel' fields to build a dependency graph, ensuring agents are only launched for tasks that are currently unblocked.

Do the agents work on different branches?

No, all agents work within the same shared epic branch. They follow strict coordination rules and use specific file patterns to minimize merge conflicts while working in parallel.

How do I monitor the progress of active agents?

You can use the companion /pm:epic-status command or view the execution-status.md file created in your epic directory for a real-time overview of active, queued, and completed streams.

What happens if an agent launch fails?

If an agent fails to start, the system provides a detailed error report for that specific stream and asks for user confirmation before proceeding with other ready tasks.

Epic Multi-Agent Orchestrator

Name: Epic Multi-Agent Orchestrator
Author: BoringHappy

byBoringHappy

0•

协作工具

Orchestrates parallel AI agents to execute complex epics and multi-task features within a shared development branch.

The pm-epic-start skill is a specialized project management tool for Claude Code that automates the transition from planning to execution for large-scale development tasks. It handles automated Git branch operations, analyzes complex task dependency graphs within your epic documentation, and intelligently launches multiple parallel sub-agents to work on different components simultaneously. By managing coordination rules and tracking real-time execution status, it allows teams to significantly accelerate feature delivery while ensuring all AI agents remain synchronized on a single source of truth.

主要功能

01Parallel sub-agent launching using the native Task tool

020 GitHub stars

03Intelligent dependency graph analysis to identify ready tasks

04Automated Git branch creation and lifecycle management

05Real-time execution tracking with auto-generated status reports

06Stream-based coordination to prevent multi-agent code conflicts

使用场景

01Managing complex refactoring projects that require simultaneous changes across different service layers

02Accelerating large feature development by distributing work across multiple concurrent AI agents

03Automating the setup and execution of multi-stage project roadmaps from Markdown-based PRDs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add boringhappy/ccpm-plugin pm-epic-start

For use in Claude.ai and ChatGPT

Download Skill