handdrawn-live-video-generator
For creators making surreal short videos that blend rough glowing hand-drawn animation with live-action spaces. Users provide a scene idea, contact object or hand, desired mood, and optional language or style constraints. The Skill clarifies the physical contact, designs continuous morphing, escape route, and delayed handheld chase movement, then writes a reusable 15-second 16:9 video prompt in the user's language. After user confirmation it recommends MiniMax H3 generation and checks contact re
Security Assessment
About handdrawn-live-video-generator
Handdrawn Live-Action Fusion Video Generator writes prompts for a finished 15-second, 16:9 short video that blends rough, glowing hand-drawn animation with live-action space. It addresses the difficulty of specifying a very particular surreal effect - a flat luminous hand-drawn entity that appears in a real room, physically contacts real hands or objects in the first 0-3 seconds, continuously morphs as one entity, escapes, and is followed by a handheld phone camera that lags slightly behind.
The skill runs a structured prompt-writing workflow: it interprets the user's request as a set of same-language constraints, invents fresh creative content each run (new space, initial form, color, morph chain, contact method, chase route, and a final space-scale transformation) rather than copying the reference prompt, and enforces a required output format with a fixed opening sentence pattern and an ordered set of time-segment paragraphs (0-3, 3-6, 6-10, 10-13, 13-15 seconds), texture, camera behavior, prohibitions, and ambient sound. It insists the crayon/chalk/pastel rough-stroke texture and non-horror, cozy tone are preserved and explicitly bans 3DCG, plush looks, smooth neon, and horror jump-scare imagery. Output language follows the dominant language of the user's input (Chinese, English, or Japanese), and it ships both English and Chinese SKILL files.
It targets creators making single-scene surreal short clips and social content. After delivering the prompt it recommends MiniMax H3 as the confirmed generation step and does not generate until the user confirms. It is not for polished CG, horror content, plush characters, or multi-scene cuts, and it requires the MiniMax Hub agent, so it is not portable to generic harnesses.
FAQ
What does the skill output?
A reusable, executable 15-second 16:9 video-generation prompt in the user's dominant input language, followed by one short next-step recommendation to generate with MiniMax H3.
Does it generate the video automatically?
No. It writes and delivers the prompt first and only recommends MiniMax H3 generation; it does not generate video, images, audio, or intermediate assets until the user explicitly confirms.
What languages does it support?
The final prompt follows the dominant language of the user's input - Chinese for Chinese, English for English, Japanese for Japanese - and the package ships both English and Chinese SKILL files.
What styles or content are prohibited?
3DCG, plush-toy looks, uniform vector lines, smooth neon, and horror imagery such as giant eyes, teeth, threatening lunges, sudden black screens, and jump scares; the tone must stay cute and non-horror.
What are its compatibility requirements?
It requires the MiniMax Hub agent (canvas workspace and MiniMax H3 generation) and is not portable to generic agent harnesses.
Install handdrawn-live-video-generator
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
minimax-ai/minimax-h3