image-qa

Community

Approve or reject images with learning feedback.

Authorredbananastudios
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill prevents wasted iterations by judging generated image outputs and converting approvals or rejections into structured learning signals that improve future results.

Core Features & Use Cases

  • Approve outcomes with evidence: Marks jobs as approved, copies winning images into selected outputs, and records winning patterns (including seeds, prompts, and LoRAs used).
  • Reject outcomes with actionable diagnosis: Categorizes failure type, records failed patterns with suggested fixes, and moves failed jobs to the failed queue.
  • Improves future generation via registries: Promotes or revokes candidate LoRAs when applicable and appends workflow success metrics to a workflow registry.

Quick Start

Ask the assistant to approve or reject the generated image results for a specific job ID, and include any notes or a failure category so the Skill can update metadata, write pattern records, and suggest the next retry strategy.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: image-qa
Download link: https://github.com/redbananastudios/ai-library/archive/main.zip#image-qa

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.