Video Prompt Reverse Tool
Paste the following prompt into your AI chat to install this skill:
Please install @user_5a849fc6/video-prompt-reverse-tool according to https://skillhub.cn/install/skillhub.md.
About this skill
Problem Addressed
Short videos usually expose only the final frames, not reusable generation parameters. Turning Kuaishou, Douyin, or local clips into AI-image-style structured prompts by hand often misses shot type, camera move, wardrobe, tone, or scene atmosphere.
How It Works
The skill follows a fixed workflow: locate or download the video by URL or local path, extract keyframes, send multiple frames to the Doubao desktop CDP channel for image explanation, then distill the response into a prompt.
- Core capability: covers shot size, camera position, camera movement, costume, color grade, composition, and scene mood.
- Key steps:
- Build a frame directory using the
f_XXX.jpgpattern, with frame numbers at multiples of 40. - Use
doubao_cdp.pyto upload keyframes from the opening, middle, and ending. - Connect through CDP to the Doubao desktop, trigger “Explain image” by scanning buttons with JS, and avoid stale refs.
- Emit a storyboard-style prompt using the template in
references/output_patterns.md.
Boundaries
It fits short clips with clear visual language, but weak for abstract material without people or explicit camerawork. Operators should watch PowerShell Chinese encoding, the CDP Tab ID, slash-style script paths, and naming mistakes that cause missed frames.
Use Cases
- Given a trending Douyin short, break down opening, middle, and ending shot size, camera angle, and movement into reusable storyboard prompts.
- For a saved Kuaishou outfit video, convert costume, color grade, composition, and scene atmosphere into structured AI-image prompts.
- With a video URL and Doubao desktop CDP setup, batch-explain keyframes and extract prompts using the documented workflow.
- When reviewing one short video, summarize shot language and mood using selected keyframes such as f_000, f_080, and f_160.
Best For
- Short-video editors who need to reverse-engineer shot size, camera angle, and movement from finished clips.
- AIGC producers who want to turn reference videos into structured prompts for image or video generation.
- Engineers automating Doubao desktop image explanation who need a stable CDP batch-click workflow.
- Script operators organizing storyboards who want costume, color, and scene atmosphere extracted from short videos.
Related Skills
Generate an interactive HTML product detail page from images and copy, with multi-product carousel, inline text editing, design controls, and PDF/JPG/PSD export support.
Enter a material name to get preview and download links for free, commercially usable video assets, with paid platform options, pricing, and search suggestions when results are limited.
Generate and validate JianYing drafts from storyboard JSON with narrated commentary, original-sound preservation, timing checks, and draft inspection.
Restyle a short video into a new visual style while preserving core elements such as characters, actions, and composition, suitable for various creative conversions like anime, illustration, ink wash, etc.