long-running-harness

Generate and maintain structured project artifacts for cross-session development.

16|Updated Nov 22, 2025

Search Tags:#workflow #long-running #engineering #multi-session #project-management #harness #AI-agent

Ready to run

npx skills add https://github.com/dapi/claude-code-marketplace --skill long-running-harness

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: long-running-harness
Source: https://github.com/dapi/claude-code-marketplace/tree/main/dev-tools/skills/long-running-harness
Command: npx skills add https://github.com/dapi/claude-code-marketplace --skill long-running-harness

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI agents typically run in isolated sessions without memory of prior work. This leads to repeated setup, loss of context between sessions, and project drift. The Long-Running Harness provides a structured orchestration pattern and artifacts to carry context, track progress, and enable cross-session development.

Core Features & Use Cases

Cross-session orchestration: Initialize and continue long-running development projects across multiple Claude sessions with persistent artifacts.
Progress tracking artifacts: Automatically generate and update artifacts like a features list and progress log to prevent context loss.
Session handoff & governance: Provide deterministic handoffs between sessions, enabling what’s next and status queries without rebuilding context.

Quick Start

Say: 'init long-running harness for Todo App' to initialize a new multi-session project and generate initial artifacts.

Frequently Asked Questions about long-running-harness

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

FAQPage Schema

How do I maintain AI agent context across multiple development sessions?▼

Cross-session context maintenance requires persistent artifacts and structured orchestration. This harness generates features.json, progress.md, and architecture.md during initialization, then applies them at the start of each new session to restore project state, track completed work, and resume development without rebuilding context.

What's the best way to coordinate long-running development projects with AI agents?▼

Long-running projects need deterministic handoffs and progress tracking. Initialize the harness with your project scope to generate structured artifacts that document features, architecture, and session progress. The harness applies these artifacts at each session start to enable seamless continuation and prevent project drift.

How do I prevent losing project context when switching between sessions?▼

Session continuity depends on persisting state across runs via structured artifacts. The harness manages multi-session development by maintaining a features list, progress log, and architecture documentation that carry forward between sessions, eliminating repeated setup and ensuring consistent project direction.

Can I use this approach for complex, multi-phase development projects?▼

Yes. The harness handles complex projects by managing feature handoffs, session orchestration, and progress governance across multiple phases. It satisfies functional and technical requirements for generating structured project artifacts and coordinating development via defined orchestration flow and allowed tools.

What artifacts does the harness generate to track project progress?▼

The harness generates three core artifacts: features.json (enumerated feature scope), progress.md (session-by-session status log), and architecture.md (structural decisions). These structured outputs persist across sessions and enable what's-next queries without rebuilding context from scratch.