iterative-runner

Automate Claude retry loops until a defined success condition is met.

1|Updated Jan 4, 2026
One-click install
npx skills add https://github.com/BarisSozen/claude --skill iterative-runner
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: iterative-runner
Source: https://github.com/BarisSozen/claude/tree/main/.claude/skills/iterative-runner
Command: npx skills add https://github.com/BarisSozen/claude --skill iterative-runner

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates Claude-based retry loops by continuously invoking Claude until a defined success condition is met.

Core Features & Use Cases

  • Persistent retry loop: Keeps invoking Claude until completion based on explicit success signals.
  • Configurable completion: Use a completion-promise or success condition to define when to stop.
  • Safety controls: Enforce max-iterations to prevent infinite loops; supports liveness checks for long-running tasks.
  • Use Case: Ideal for TDD loops, flaky experiments, overnight builds, or any scenario requiring repeated iterations until the desired outcome is achieved.

Quick Start

iterative-runner "Build feature X" --completion-promise "DONE" --max-iterations 30

Frequently Asked Questions about iterative-runner

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

FAQPage Schema
How do I automate Claude to keep retrying a prompt until a test passes?

Automate Claude retry loops by continuously invoking Claude with a given prompt until a defined success condition is met. You specify a completion-promise and max-iterations to control when the loop stops.

What is a completion-promise in automated retry loops?

A completion-promise is an explicit success signal that defines when a retry loop stops invoking Claude. When the output matches this condition, the loop terminates successfully.

Can I use iterative retry loops for TDD workflows and overnight builds?

Retry loops are ideal for TDD workflows, flaky experiments, and overnight builds. They repeatedly iterate through Claude prompts until the desired outcome is achieved.

How do I prevent infinite loops when automating Claude iterations?

Prevent infinite loops by enforcing max-iterations to cap the number of attempts. The system also supports liveness checks for long-running tasks to ensure safe execution.

Does the retry loop support persistent prompts for repeated Claude invocations?

The retry loop supports persistent prompts, maintaining the given prompt context across continuous Claude invocations until the defined success condition is finally met.