How are optimization results evaluated?

Results are evaluated using fitness functions, such as 'skill_quality' or LLM-based 'default' evaluation, which check for clarity, completeness, and structural integrity.

What does the /optimize command do?

It analyzes a targeted skill file and uses an LLM-driven pipeline to rewrite the content for better performance based on defined fitness criteria and past context.

What is the purpose of the --dry-run flag?

The dry-run flag performs a structural test of the optimization process without calling the LLM or modifying the content, ensuring the pipeline is configured correctly.

What is the difference between global and project scopes?

Global scope optimizes skills located in your home directory for general use, whereas project scope includes specific codebase context to tailor rules for a particular repository.

Can I revert changes if I'm not happy with the optimization?

Yes, the skill includes a --restore flag that allows you to immediately roll back to a previous version from the automatically created backup.

Genetic Prompt Optimizer

Name: Genetic Prompt Optimizer
Author: min-sys

bymin-sys

•

Herramientas para Desarrolladores

Optimizes Claude Code skills and rules through an autonomous evolutionary pipeline using LLM-driven direct patching.

The Genetic Prompt Optimizer is a sophisticated tool designed to refine and evolve Claude Code skills (SKILL.md) by applying iterative improvements based on performance feedback and specific context. By utilizing genetic optimization concepts and fitness functions, it enables Claude to autonomously rewrite its own operational rules to achieve higher levels of clarity, completeness, and practical utility. This skill automates the prompt engineering process, allowing for systematic upgrades to AI behavior through a structured workflow that includes error-guided improvements and human-in-the-loop verification.

Características Principales

01Autonomous direct patching and optimization of SKILL.md files

02Integrated version management with dry-run, restore, and history logging

03Multiple optimization modes including auto, error-guided, and LLM-improve

041 GitHub stars

05Customizable fitness functions to evaluate skill quality and clarity

06Support for both global and project-specific skill scopes

Casos de Uso

01Improving the clarity and consistency of project-specific documentation and guidelines

02Refining complex coding rules based on failure logs and execution context

03Evolving general-purpose global skills to perform better across diverse development environments

Características Principales

01Autonomous direct patching and optimization of SKILL.md files

02Integrated version management with dry-run, restore, and history logging

03Multiple optimization modes including auto, error-guided, and LLM-improve

041 GitHub stars

05Customizable fitness functions to evaluate skill quality and clarity

06Support for both global and project-specific skill scopes

Casos de Uso

01Improving the clarity and consistency of project-specific documentation and guidelines

02Refining complex coding rules based on failure logs and execution context

03Evolving general-purpose global skills to perform better across diverse development environments