🎁 Get the FREE AI Skills Starter Guide β€” Subscribe β†’
BytesAgainBytesAgain
πŸ¦€ ClawHub

Zouroboros Autoloop

by @marlandoj

Autonomous optimization loop inspired by Andrej Karpathy's autoresearch: edit, experiment, measure, keep or revert. Best for any task with a numeric metric.

Versionv1.0.0
⚑ When to Use
TriggerAction
- **Prompt optimization** β€” Tune prompts to maximize a quality metric
- **Site performance** β€” Reduce load time, optimize bundle size
- **Model fine-tuning** β€” Iterate on hyperparameters with measurable output
πŸ’‘ Examples

autoloop --program  [--executor ] [--resume] [--dry-run]

  • --program β€” Path to your program.md specification (required)
  • --executor β€” Shell command that reads a prompt from stdin and outputs a response
  • - OpenClaw: "openclaw ask" - Claude Code: "claude --print" - Any LLM CLI that reads stdin
  • --resume β€” Resume from an existing autoloop branch
  • --dry-run β€” Validate program.md without running
  • βš™οΈ Configuration

    npm install
    

    πŸ“‹ Tips & Best Practices

    Focus on risk-adjusted returns. Avoid overfitting to recent data. ```

    πŸ”’ Constraints

  • Time budget per run: 60
  • Max experiments: 50
  • Max duration: 4
  • Max cost: 5
  • View on ClawHub
    TERMINAL
    clawhub install zouroboros-autoloop

    πŸ§ͺ Use this skill with your agent

    Most visitors already have an agent. Pick your environment, install or copy the workflow, then run the smoke-test prompt above.

    πŸ” Can't find the right skill?

    Search 60,000+ AI agent skills β€” free, no login needed.

    Search Skills β†’