hiddink-harness-loop

Implement a SubagentStop prompt hook to resume workflow execution after subagent completion.

1|Updated May 21, 2026
One-click install
npx skills add https://github.com/hiddink-ai/hiddink-harness --skill hiddink-harness-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hiddink-harness-loop
Source: https://github.com/hiddink-ai/hiddink-harness/tree/main/templates/skills/omcustom-loop
Command: npx skills add https://github.com/hiddink-ai/hiddink-harness --skill hiddink-harness-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill prevents AI coding sessions from idling when background subagents finish their tasks, ensuring that multi-step workflows continue processing without requiring constant manual user intervention.

Core Features & Use Cases

  • SubagentStop Hook: Automatically triggers a check for pending workflow steps whenever a background subagent completes its execution.
  • Safety Guardrails: Includes a 3-step limit on auto-continuation and a stuck-detector to prevent infinite loops or excessive cost accumulation.
  • Use Case: When running a complex refactoring task that spawns multiple background subagents, this skill ensures the orchestrator immediately picks up the next task in the queue rather than waiting for the user to manually prompt it.

Quick Start

Invoke the hiddink-harness-loop skill to verify the current status of the auto-continuation hooks in your active session.

Frequently Asked Questions about hiddink-harness-loop

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

FAQPage Schema
How do I prevent my Claude Code session from idling when background subagents finish?

To prevent session idling when background subagents finish, you can use a SubagentStop prompt hook that automatically resumes workflow execution upon subagent completion, checking for pending steps without manual intervention.

What is a SubagentStop hook and how does it manage background task transitions?

A SubagentStop hook is a prompt hook that automatically triggers a check for pending workflow steps whenever a background subagent completes its execution, managing background task transitions within the Claude Code orchestration environment.

How do I automatically continue multi-step workflows after a subagent completes?

You can automatically continue multi-step workflows by implementing an auto-continuation hook that validates pending steps and resumes the orchestrator immediately, ensuring the next task in the queue is picked up automatically.

Can I limit autonomous workflow progression to prevent infinite loops and excessive costs?

Yes, you can enforce safety limits on autonomous progression using a 3-step consecutive execution cap and a stuck-detector integration, which prevent infinite loops and control cost accumulation during auto-continuation.

Does auto-continuation work with complex refactoring tasks that spawn multiple background subagents?

Yes, auto-continuation works with complex refactoring tasks by ensuring the orchestrator immediately picks up the next task in the queue when multiple background subagents complete, rather than waiting for manual user prompts.

When should I avoid using auto-continuation hooks in my orchestration environment?

You should avoid using auto-continuation hooks when you need strict manual oversight of each workflow step, as the automatic progression bypasses manual validation and relies on built-in safety caps to prevent runaway execution.