Is there a limit to how many times the prompt is optimized?

The skill is designed to converge rather than cap, meaning it will continue to run until the prompt-reviewer identifies no further issues.

Which other skills are required for this to work?

This skill acts as an orchestrator and requires the prompt-reviewer and prompt-engineering skills to be available in your environment.

Does this skill overwrite my existing prompt files?

No, it works on a temporary file in /tmp/ during the optimization iterations and only applies the final, optimized version to your original file once all issues are resolved.

What happens if an issue requires human clarification?

If the optimization process encounters a 'NEEDS_USER_INPUT' status, it will pause to ask you for clarification or options before proceeding with the next iteration.

Auto-Optimize Prompt

Name: Auto-Optimize Prompt
Author: doodledood

bydoodledood

•

生産性とワークフロー

Iteratively refines and fixes prompt issues through an automated review-and-repair loop until complete convergence is achieved.

The Auto-Optimize Prompt skill streamlines the prompt engineering process by automating the tedious cycle of reviewing, debugging, and improving LLM instructions. By chaining specialized review and engineering capabilities, it identifies structural weaknesses and ambiguities, asks for user clarification when necessary, and applies fixes to a temporary working copy. This ensures your final prompt is production-ready and optimized for performance without requiring manual trial and error or multiple rounds of human editing.

主な機能

01Atomic output logic that preserves the original file until the final version is ready

02Delegates specialized tasks to prompt-reviewer and prompt-engineering tools

0363 GitHub stars

04Automated iterative refinement loop that runs until all issues are resolved

05Comprehensive reporting on iterations, issues fixed, and summary of changes

06User-in-the-loop resolution for complex ambiguities requiring human context

ユースケース

01Polishing complex system prompts for AI agents to ensure high-quality outputs

02Debugging underperforming prompts that currently produce inconsistent or incorrect results

03Standardizing prompt quality across a codebase by fixing structural issues automatically

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doodledood/manifest-dev auto-optimize-prompt

For use in Claude.ai and ChatGPT

主な機能

01Atomic output logic that preserves the original file until the final version is ready

02Delegates specialized tasks to prompt-reviewer and prompt-engineering tools

0363 GitHub stars

04Automated iterative refinement loop that runs until all issues are resolved

05Comprehensive reporting on iterations, issues fixed, and summary of changes

06User-in-the-loop resolution for complex ambiguities requiring human context

ユースケース

01Polishing complex system prompts for AI agents to ensure high-quality outputs

02Debugging underperforming prompts that currently produce inconsistent or incorrect results

03Standardizing prompt quality across a codebase by fixing structural issues automatically

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doodledood/manifest-dev auto-optimize-prompt

For use in Claude.ai and ChatGPT