narrative-scene-qa

Name: narrative-scene-qa
Availability: InStock
Author: ufukdegen

Compare rendered narrative images against canonical beat intent and return JSON verdicts.

178|18|Updated May 3, 2026

Search Tags:#narrative #qa #vision #continuity #canon #fidelity #image-eval

Ready to run

npx skills add https://github.com/UfukNode/Noustiny --skill narrative-scene-qa

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: narrative-scene-qa
Source: https://github.com/UfukNode/Noustiny/tree/main/hermes-additions/skills/creative/narrative-scene-qa
Command: npx skills add https://github.com/UfukNode/Noustiny --skill narrative-scene-qa

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Compare a rendered narrative image against the beat's intent (title + body + prior canon) and flag faithfulness failures so the image pipeline can re-render with correction. Returns strict JSON {approved, score, issues, fix_hint}. No prose, no tool calls.

Core Features & Use Cases

Post-render verification of image alignment to canonical state and beat intent.
Deterministic JSON verdicts for automated downstream processing.
Safe, prose-free evaluation suitable for integration into rendering pipelines.

Quick Start

Provide the rendered image prompt and canonical beat context to validate fidelity against the beat's intent.

Frequently Asked Questions about narrative-scene-qa

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

FAQPage Schema

How do I verify narrative image fidelity after rendering?▼

You can verify narrative-image fidelity by comparing the rendered output against the beat's title, body, and prior canon. This flags faithfulness failures and returns a deterministic JSON payload for automated downstream correction.

What is narrative continuity checking for rendered images?▼

Narrative continuity checking is a post-render evaluation that compares visual output against canonical beat data. It flags faithfulness failures between the image and the scene intent to ensure strict visual alignment.

How do I get automated JSON feedback for image pipeline errors?▼

You get automated JSON feedback by validating the rendered image prompt with canonical beat context. The evaluation returns a strict JSON payload containing approved status, score, issues, and fix_hint without any prose.

Can I use this for automated post-render image evaluation?▼

Yes, you can use this for automated post-render image evaluation because it operates safely without prose or tool calls. It provides deterministic JSON verdicts specifically designed for integration into automated rendering pipelines.

What is the best way to flag canon inconsistencies in generated scenes?▼

The best way to flag canon inconsistencies is to compare the rendered narrative image against the canonical beat data. This process identifies faithfulness failures and returns a structured JSON verdict with a fix_hint.

Why does my narrative image evaluation return a strict JSON payload?▼

Your narrative image evaluation returns a strict JSON payload to ensure safe, prose-free evaluation suitable for automated downstream processing. This deterministic format includes approved, score, issues, and fix_hint fields for automated re-rendering.