long-running-agent

Coordinate autonomous end-to-end software project delivery with persistent state files.

48|4|Updated Mar 4, 2026

Search Tags:#worktrees #subagents #orchestration #autonomous development #milestone planning #project delivery

Ready to run

npx skills add https://github.com/jarrodwatts/long-running-agent-skill --skill long-running-agent-jarrodwatts

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: long-running-agent
Source: https://github.com/jarrodwatts/long-running-agent-skill/tree/main
Command: npx skills add https://github.com/jarrodwatts/long-running-agent-skill --skill long-running-agent-jarrodwatts

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you manage large software projects from kickoff to completion when the work spans many steps, many files, and long stretches of autonomous execution.

Core Features & Use Cases

End-to-end orchestration: Break a broad project goal into milestones, tasks, and verification steps.
Parallel implementation: Delegate complex work to isolated subagents so multiple parts of a project can move forward at once.
Persistent state management: Use .agent/ files to track goals, plans, decisions, progress, and review feedback across long sessions.
Quality enforcement: Run review cycles, fix issues, and keep the project aligned with architecture and testing standards.
Use case: A team asks for a full product implementation, and this Skill organizes discovery, planning, execution, review, and completion without constant supervision.

Quick Start

Use the long-running-agent skill to plan, execute, review, and finish this software project end to end with autonomous milestone tracking.

Frequently Asked Questions about long-running-agent

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

FAQPage Schema

How do I run autonomous development for a large software project end to end?▼

Autonomous end-to-end software project delivery works by breaking broad goals into milestones, delegating tasks to isolated worktree subagents, and tracking progress in persistent .agent state files without repeated user intervention.

What is the best way to coordinate parallel tasks in a long-running agent workflow?▼

Parallel task coordination uses isolated worktree subagents to execute multiple project parts simultaneously, while milestone planning and persistent state files keep the overall execution deterministic and recoverable.

How does milestone planning work for multi-milestone project implementations?▼

Milestone planning breaks a broad project goal into discrete tasks and verification steps, using .agent files to track goals, decisions, and review feedback across long autonomous sessions.

Can I use isolated worktree subagents for greenfield builds without constant supervision?▼

Yes, isolated worktree subagents support greenfield builds by handling parallel implementation autonomously, relying on persistent state files and review cycles to maintain quality without constant supervision.

Do I need persistent state files to manage long-running maintenance sessions?▼

Persistent .agent state files are required to track progress, decisions, and review feedback during long-running maintenance sessions, ensuring execution remains deterministic and recoverable across extended periods.

When should I not use autonomous orchestration for software project delivery?▼

Avoid autonomous orchestration for tasks requiring frequent manual intervention or continuous user feedback, as this approach relies on isolated execution, milestone planning, and verification cycles to minimize supervision.