Why does this skill recommend 4-second video clips?

Sora follows instructions most accurately in 4-second segments. Longer durations are better suited for simple, ambient scenes with less complex motion.

What is the 'Golden Rule' for reference images in Sora?

The Golden Rule dictates that when using a reference image, you should describe motion and action only. Do not re-describe the character or setting already visible in the image.

How should I structure a video prompt for the best results?

Follow the Sanzaru Prompt Anatomy: Start with Style, followed by Scene, then Camera movement, Action in beats, and finally Lighting & Color.

Which image generation tool is recommended for simple tasks?

The 'generate_image' tool is recommended for most cases because it is synchronous and returns the result immediately without the need for status polling.

Sanzaru Multimodal Prompting Guide

Name: Sanzaru Multimodal Prompting Guide
Author: TJC-LP

byTJC-LP

•

Content Management

Guides the creation of high-quality AI video, image, and audio assets using optimized prompting patterns for the Sanzaru MCP server.

The Sanzaru Prompt Guidance skill equips Claude with specialized knowledge for generating multimodal content through the OpenAI API, specifically focusing on Sora video and GPT-Image models. It enforces the 'Golden Rule' of reference images—describing motion only—to prevent redundant prompting and ensures Claude follows a structured 'Sora Prompt Anatomy' involving style, scene, camera, action, and lighting. This skill is essential for users looking to automate cinematic video production, iterative image refinement, and high-quality text-to-speech workflows while maintaining strict adherence to model-specific limitations and polling patterns.

Key Features

015 GitHub stars

02Motion-only prompting logic for reference image consistency

03Optimized model selection between Sora-2 and Sora-2-Pro

04Structured Sora Prompt Anatomy for cinematic video results

05Standardized async polling patterns for video and image jobs

06Media dimension validation and preparation guidance

Use Cases

01Generating professional AI video clips with precise camera movements

02Automating complex multimodal workflows involving TTS and video remixing

03Orchestrating iterative image refinement chains using the Responses API

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tjc-lp/sanzaru prompt-guidance

For use in Claude.ai and ChatGPT

Download Skill