aif-loop

Manage iterative artifact refinement loops across PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE phases.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/ichinya/idshka --skill aif-loop-ichinya
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: aif-loop
Source: https://github.com/ichinya/idshka/tree/main/.codex/skills/aif-loop
Command: npx skills add https://github.com/ichinya/idshka --skill aif-loop-ichinya

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Reflex loops are complex coordination tasks that help improve an artifact by iterative planning, production, evaluation, critique, and refinement until acceptance criteria are met.

Core Features & Use Cases

  • Orchestrates multi-phase cycles (PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, REFINE) with strict phase contracts
  • Supports parallel execution where possible to speed up iteration
  • Maintains persistent state across sessions for resumable workflows

Quick Start

Run an automated reflex loop to progressively refine an artifact across PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE.

Frequently Asked Questions about aif-loop

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

FAQPage Schema
How do I automate iterative refinement for artifacts until they meet acceptance criteria?

An automated reflex loop coordinates PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE phases to progressively improve artifacts. It persists state between iterations and enforces configurable stop conditions for reproducible iterative refinement.

How does a reflex loop manage state across multiple evaluation cycles?

Reflex loops maintain persistent state across iterations by recording artifacts, checks, and history during each phase. This persistent state enables reproducibility and supports resuming workflows across different sessions without losing prior evaluation progress.

Can I run evaluation and critique phases in parallel to speed up artifact iteration?

Yes, the workflow supports parallel execution where possible during the iterative phases. Running compatible operations concurrently reduces total cycle time while maintaining strict phase contracts across the PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE stages.

What's the best way to resume an interrupted artifact refinement workflow?

You can resume an interrupted artifact refinement workflow by leveraging the persistent state recorded across sessions. The loop stores history, checks, and artifacts at each phase boundary, allowing you to restart from the last completed step without re-running the entire pipeline.

How do I configure stop conditions for an automated evaluation and critique loop?

You can configure stop conditions within the reflex loop to define when the iterative refinement should terminate. These conditions act as acceptance criteria evaluated during the EVALUATE phase, ensuring the loop only stops when artifacts meet your specified quality thresholds.

Why use strict phase contracts in an automated artifact refinement workflow?

Strict phase contracts ensure that each stage of the reflex loop—PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, and REFINE—produces predictable, reproducible outputs. This structural consistency guarantees that artifacts, checks, and history are reliably recorded across all iterations.