Loop

What Is Loop?

The "Loop" skill is an autonomous scheduling utility designed for Claude Code, enabling users to execute experiments or tasks at user-defined intervals without manual intervention. By leveraging CronCreate for precise scheduling, Loop allows for seamless automation of recurring experiments—whether they’re rapid-fire checks every 10 minutes or long-term studies running monthly. The skill is invoked using the /ar:loop command and integrates with the Claude Autoresearch Agent framework, making it a valuable addition for automating research and engineering workflows.

Why Use Loop?

Research and engineering often require repetitive, scheduled execution of scripts, tests, or data collection routines. Traditionally, this necessitates manual scheduling with external tools or repeated manual triggers, which can introduce errors, waste time, and reduce productivity. Loop addresses these challenges by embedding a robust, user-friendly scheduling interface directly within the Claude environment.

Key use cases include:

• Automated Performance Monitoring: Regularly benchmark APIs or services at fixed intervals.
• Longitudinal Experimentation: Run experiments over days, weeks, or months to track changes over time.
• Continuous Integration: Schedule code quality checks or test suites to run automatically.
• Background Data Collection: Automate data harvesting and preprocessing tasks.

By eliminating manual oversight, Loop frees up valuable time for engineers and researchers to focus on analysis and innovation.

How to Get Started

Getting started with Loop is straightforward. First, ensure that the Loop skill is installed and accessible within your Claude Code workspace. The primary command is /ar:loop, followed by the name of the experiment and, optionally, the desired interval.

Basic Usage

Start a new loop for a specific experiment:

/ar:loop engineering/api-speed

If you omit the interval, Loop will prompt you to select one interactively. To specify the interval directly, provide it as an argument:

/ar:loop engineering/api-speed 10m      # Every 10 minutes
/ar:loop engineering/api-speed 1h       # Every hour
/ar:loop engineering/api-speed daily    # Daily at ~9am
/ar:loop engineering/api-speed weekly   # Weekly on Monday ~9am
/ar:loop engineering/api-speed monthly  # Monthly on 1st ~9am

To stop an active loop:

/ar:loop stop engineering/api-speed

Interactive Interval Selection

If no interval is provided, Loop will present you with preset options:

Select loop interval:
  1. Every 10 minutes  (rapid — stay and watch)
  2. Every hour        (background — check back later)
  3. Daily at ~9am     (overnight experiments)
  4. Weekly on Monday  (long-running experiments)
  5. Monthly on 1st    (slow experiments)

Key Features

Loop is designed for flexibility, reliability, and ease of use. Its main features include:

• User-Selected Scheduling: Supports intervals of 10 minutes, hourly, daily, weekly, and monthly.
• Cron Integration: Maps human-friendly intervals to precise cron expressions for accurate timing.
• Interactive Experiment Selection: If no experiment is specified, Loop prompts the user to select from available experiments.
• Automated Execution: Once configured, Loop autonomously triggers the experiment at the selected interval.
• Simple Stop Command: Halting an experiment loop is as easy as issuing the stop command with the relevant experiment name.

Cron Expression Mapping

Loop transparently handles the conversion from common interval names to cron expressions. For reference:

This mapping ensures that tasks are triggered at consistent and predictable times.

Practical Example

Suppose you want to benchmark an API every hour:

/ar:loop engineering/api-speed 1h

Loop will schedule the engineering/api-speed experiment to execute every hour automatically. If you later wish to stop this loop:

/ar:loop stop engineering/api-speed

Best Practices

To maximize the effectiveness of Loop in your workflows:

• Choose Appropriate Intervals: Select an interval that matches the frequency needed for your experiment. Avoid overly frequent intervals for resource-intensive tasks.
• Monitor Output: Regularly check the results of your scheduled experiments, especially for long-running loops.
• Name Experiments Clearly: Use descriptive experiment names to avoid confusion when managing multiple loops.
• Combine with Notification Systems: Integrate with alerting or reporting tools to receive updates on experiment outcomes.
• Review Scheduled Loops: Periodically review active loops to ensure relevance and prevent unnecessary resource consumption.

Important Notes

• Resource Management: Running multiple frequent loops can consume significant computational resources. Monitor system load and adjust intervals as necessary.
• Time Zone Considerations: Scheduled times are based on the server’s time zone. Ensure alignment with your operational context.
• Experiment Availability: If the target experiment is deleted or renamed, scheduled loops may fail silently. Keep experiments up-to-date and verify their status.
• CronCreate Dependency: Loop relies on CronCreate for all scheduling. Ensure this dependency is properly configured and running for Loop to function.
• Security: Only authorized users should be able to schedule or stop loops, particularly in shared or sensitive environments.

By following these guidelines and leveraging Loop’s automation capabilities, you can streamline repetitive experiment cycles, improve consistency, and enhance productivity in research and development workflows.