Does it support custom quality rules?

Absolutely. Users can define specific success criteria, mandatory checks, and quality thresholds for both initial (Phase A) and final (Phase B) evaluations.

How does the parallel execution model work?

It utilizes the Task tool to run artifact production and test preparation simultaneously, significantly speeding up the overall iteration cycle.

Can I resume an interrupted refinement loop?

Yes, the skill maintains persistent state in .ai-factory files, allowing you to resume, stop, or check the status of any loop at any time.

What is the primary purpose of the aif-loop skill?

It automates a structured 'generate-critique-refine' cycle to ensure high-quality output through multiple iterations and strict evaluation phases.

Reflex Iteration Workflow

Name: Reflex Iteration Workflow
Author: hightemp

byhightemp

0•

Herramientas para Desarrolladores

Executes a multi-phase reflex loop to iteratively generate, evaluate, and refine high-quality artifacts using strict quality gates.

The aif-loop skill implements a sophisticated 'generate-critique-refine' workflow designed to improve the quality and reliability of Claude's outputs. By following a structured six-phase process—PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE—it ensures that every artifact, whether code, documentation, or API specifications, meets specific project requirements and quality thresholds. This skill is particularly useful for complex tasks that require rigorous validation, parallel execution of tests, and persistent state management across multiple development sessions.

Características Principales

01Tiered quality gates with Phase A/B success thresholds

02Parallel execution of artifact production and test preparation

03Strict multi-iteration Reflex Loop (PLAN to REFINE)

040 GitHub stars

05Persistent state management via structured JSON and Markdown files

06Customizable evaluation rules and mandatory check scripts

Casos de Uso

01Generating production-grade OpenAPI specifications with automated schema validation

02Developing complex logic modules that require iterative unit testing and code refinement

03Drafting technical documentation that must adhere to strict project-specific style guides

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hightemp/openrouter-chat-app-vue aif-loop

For use in Claude.ai and ChatGPT

Características Principales

01Tiered quality gates with Phase A/B success thresholds

02Parallel execution of artifact production and test preparation

03Strict multi-iteration Reflex Loop (PLAN to REFINE)

040 GitHub stars

05Persistent state management via structured JSON and Markdown files

06Customizable evaluation rules and mandatory check scripts

Casos de Uso

01Generating production-grade OpenAPI specifications with automated schema validation

02Developing complex logic modules that require iterative unit testing and code refinement

03Drafting technical documentation that must adhere to strict project-specific style guides

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hightemp/openrouter-chat-app-vue aif-loop

For use in Claude.ai and ChatGPT