What are the different operating modes available?

Project Harness offers Supervised mode (human review after every task), Semi-Auto (review at story/epic boundaries), and Autonomous (review only at major epic milestones).

How does Project Harness integrate with task management?

It uses the ohno MCP plugin to track tasks, manage project state, and provide a visual Kanban board. This allows the AI to automatically fetch the next task, update progress, and log session notes.

What is the benefit of using subagents in Project Harness?

Subagents provide a fresh context for every individual task, preventing the performance degradation, accumulated assumptions, and token limit issues that often occur in long-running manual coding sessions.

Do I need to manually commit my code when using this skill?

No, the skill's integrated hook system automatically handles git commits, synchronization, and basic testing based on your selected operating mode and lifecycle points.

Project Harness

Name: Project Harness
Author: srstomp

bysrstomp

•

Productivity & Workflow

Orchestrates complex, long-running AI development sessions with configurable human oversight and structured task management.

About

Project Harness provides a robust framework for managing autonomous or semi-autonomous development workflows within Claude Code. By integrating with the ohno task management system, it breaks down large projects into manageable subagent-driven tasks, ensuring fresh context for every execution and preventing session degradation. It offers flexible operation modes ranging from fully supervised to autonomous, allowing developers to maintain control through strategic checkpoints and automated hooks for syncing, testing, and version control. This skill is essential for developers looking to move beyond single-prompt coding and into reliable, multi-session project implementation.

Key Features

Integrated project management using the ohno MCP and visual Kanban system
Configurable operating modes (Supervised, Semi-Auto, Autonomous) for flexible human control
2 GitHub stars
Automated lifecycle hooks for reliable git commits, syncing, and state management
Intelligent skill routing to dispatch specialized domain knowledge for specific tasks
Task-based subagent execution for fresh context and reduced token bloat

Use Cases

Standardizing AI development workflows across a team using shared project context and structured protocols
Executing autonomous development loops with safety checkpoints for human review and course correction
Managing multi-day feature implementations that require consistent progress tracking across sessions

About

Key Features

Integrated project management using the ohno MCP and visual Kanban system
Configurable operating modes (Supervised, Semi-Auto, Autonomous) for flexible human control
2 GitHub stars
Automated lifecycle hooks for reliable git commits, syncing, and state management
Intelligent skill routing to dispatch specialized domain knowledge for specific tasks
Task-based subagent execution for fresh context and reduced token bloat

Use Cases

Standardizing AI development workflows across a team using shared project context and structured protocols
Executing autonomous development loops with safety checkpoints for human review and course correction
Managing multi-day feature implementations that require consistent progress tracking across sessions