refine

Iteratively refine artifacts toward a measurable goal using a propose-measure-keep-or-revert loop.

Updated May 27, 2026
One-click install
npx skills add https://github.com/jiahknee5/rapid-workflow --skill refine-jiahknee5
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: refine
Source: https://github.com/jiahknee5/rapid-workflow/tree/main/skills/refine
Command: npx skills add https://github.com/jiahknee5/rapid-workflow --skill refine-jiahknee5

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill provides a systematic approach to iteratively refine artifacts like code, design, copy, config, or prompts towards a defined, measurable goal.

Core Features & Use Cases

  • Goal-directed refinement: Propose variations, measure against a goal, and iterate until convergence.
  • Keep-or-revert mechanism: Ensures artifacts do not regress by reverting to the previous version if no improvement is found.
  • Use Case: Refine a design mockup to meet a desirability bar, refine code to pass a performance metric, or refine a prompt to achieve a desired task outcome.

Quick Start

Run the refine skill with the artifact, goal, and evaluation criteria to start the refinement loop.

Frequently Asked Questions about refine

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

FAQPage Schema
How do I iteratively refine code or design artifacts towards a measurable goal?

To iteratively refine artifacts towards a measurable goal, use a propose-measure-keep-or-revert loop that suggests variations, evaluates them against your criteria, and keeps the improved versions while reverting regressions.

What is a keep-or-revert loop for artifact refinement?

A keep-or-revert loop for artifact refinement proposes changes, measures them against a defined goal, and permanently keeps improvements while automatically reverting to the previous version if no measurable progress is found.

How do I evaluate and improve prompt performance through iteration?

To evaluate and improve prompt performance through iteration, provide the prompt artifact, a desired task outcome goal, and evaluation criteria to trigger an automated refinement loop that converges on the target metric.

Can I use automated iteration to refine configuration files for a specific metric?

Yes, you can use automated iteration to refine configuration files by providing the config artifact, your target metric, and evaluation criteria to systematically propose variations and keep only the changes that improve the measurement.

What inputs are required to start goal-directed refinement on an artifact?

Goal-directed refinement requires three inputs: the initial artifact you want to improve, the measurable goal you are targeting, and the evaluation criteria used to measure progress and determine whether to keep or revert changes.

Does iterative refinement work for both code optimization and copywriting?

Yes, iterative refinement works for both code optimization and copywriting by supporting various modes for different artifact types, allowing you to apply the same propose-measure-keep-or-revert evaluation loop across diverse formats.