skill-evolver

Orchestrates a planning-to-evolution loop for improving skills and workflows with testable evidence.

Updated Mar 13, 2026

Search Tags:#testing #workflow automation #prompt engineering #validation #iterative improvement #skill evolution

Ready to run

npx skills add https://github.com/jf3096/personal-configs --skill skill-evolver-jf3096

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-evolver
Source: https://github.com/jf3096/personal-configs/tree/main/.codex/skills/skill-evolver
Command: npx skills add https://github.com/jf3096/personal-configs --skill skill-evolver-jf3096

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the challenge of ad-hoc prompt tuning by establishing a structured, evidence-based iteration loop for continuous improvement of skills and workflows.

Core Features & Use Cases

Rehearsal-to-Evolution Loop: Guides users through distinct stages of planning, execution, comparison, and evolution.
Evidence-Driven Improvement: Mandates concrete evidence for changes, including positive, negative, and adversarial test cases.
Use Case: When a deployed skill is underperforming or needs to be adapted to new requirements, this skill can be used to systematically refine its prompts, rules, or tests, ensuring improvements are measurable and robust.

Quick Start

Use the skill-evolver to start a new evolution cycle for the task "Convert compliance-review-demo.js to python" using the provided script.

Frequently Asked Questions about skill-evolver

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

FAQPage Schema

How do I iteratively improve prompt engineering workflows with evidence?▼

Iterative prompt improvement requires a rehearsal-to-evolution loop that validates changes through planning, execution, comparison, and evidence-backed evolution using positive, negative, and adversarial test cases.

What is the best way to validate workflow automation changes before deployment?▼

Validating workflow automation changes is best achieved by running deterministic task execution scripts and comparing results against positive, negative, and adversarial test cases to ensure measurable robustness.

How does an evidence-driven evolution loop work for refining skills?▼

An evidence-driven evolution loop works by guiding users through structured stages of planning, execution, and comparison, mandating concrete test evidence to adapt prompts and rules to new requirements.

Do I need scripts to automate testing for skill evolution cycles?▼

Yes, you need scripts to support deterministic task execution within the skill evolution cycle, ensuring that workflow improvements are validated consistently and automatically across test cases.

When should I use adversarial test cases in prompt tuning?▼

You should use adversarial test cases in prompt tuning when a deployed skill is underperforming, ensuring that evidence-backed changes to prompts and rules remain robust against edge-case failures.

Can I use references for in-depth information retrieval during skill evolution?▼

Yes, references support in-depth information retrieval during the skill evolution loop, providing the necessary context to systematically refine prompts, rules, and tests during the comparison stage.