harness-loop

Coordinate long-running task loops with fresh context and controlled termination.

39|46|Updated May 29, 2026
One-click install
npx skills add https://github.com/maxwell2732/claudecode-research-harness-workflow --skill harness-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: harness-loop
Source: https://github.com/maxwell2732/claudecode-research-harness-workflow/tree/main/skills/harness-loop
Command: npx skills add https://github.com/maxwell2732/claudecode-research-harness-workflow --skill harness-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Coordinating long-running task loops within Claude Code Harness using /loop and ScheduleWakeup to re-enter with fresh context on each wake-up.

Core Features & Use Cases

  • Integrates /loop, ScheduleWakeup, and harness-work to create repeatable task cycles with isolated contexts.
  • Manages plan/contractReview/advisor steps across cycles to preserve traceability and reproducibility.
  • Suitable for orchestrating complex experiments, simulations, and iterative data-processing pipelines where auditable execution is required.

Quick Start

Launch a long-running loop using /harness-loop with an initial Plans.md to begin iterative, auditable tasks.

Frequently Asked Questions about harness-loop

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

FAQPage Schema
How do I run long-running task loops with fresh context in Claude Code?

Long-running task loops with fresh context are coordinated using /loop and ScheduleWakeup to re-enter execution cycles, preserving state across iterations while isolating context for each wake-up.

What is the best way to automate iterative task cycles while maintaining memory resume?

Automated iterative task cycles with memory resume are managed by orchestrating sub-skills across cycles, preserving plan adherence and traceability through advisor and reviewer hooks.

How do I control termination for repeated task cycles based on completeness or plateau?

Controlled termination for repeated task cycles is achieved by evaluating cycles, plateau detection, or task completeness, stopping the loop automatically when defined conditions are met.

Can I orchestrate complex simulations with auditable execution using a plan management loop?

Complex simulations with auditable execution are supported by managing plan, contract review, and advisor steps across cycles, ensuring reproducibility and traceability for iterative data-processing pipelines.

How do I start an iterative task loop using an initial plan file?

Iterative task loops are started by launching the loop command with an initial Plans.md file, which provides the baseline plan for orchestrating auditable, repeatable task cycles.

Do I need any external dependencies to coordinate re-entry with fresh context?

No external dependencies are required to coordinate re-entry with fresh context; the loop orchestration relies entirely on built-in harness commands and scheduling mechanisms.