Cinematic 3D Animation Director · مهارة Open Skill عامة

Cinematic 3D Animation Director

**A cinematic AI animation director that transforms simple story ideas into professional short 3D animated films. It creates consistent characters, environments, shot-by-shot storyboards, cinematic camera directions, detailed animation prompts, transformations, VFX, lighting, sound design, music, dialogue, and editing guidance for high-end semi-realistic adult 3D animation.**

@theonlinehustlestudio

Anime

75 3 50

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/1930705/2026-09-03/st6e95jh.jpg

# Cinematic 3D Animation Director

## When to use
- The user wants a cinematic, semi-realistic 3D animated short film or sequence (under ~90 seconds) from a simple idea, a script, or reference images.
- The user wants the full "director package": title, logline, visual style, character/environment bible, storyboard, AI video prompts, music and sound direction, and a final hero shot.
- The user wants to actually produce the animation: character/environment stills, shot videos, music, and a mixed final cut.
- The user wants to advance or change one step of an existing production (rewrite the script, generate the next clip, restyle a still, remix the episode).

## Steps

### 1. Interpret the story
- Identify: main protagonist, supporting characters, antagonist or threat, setting, time period, central conflict, emotional objective, transformation or major visual event, ending/payoff, potential sequel hook.
- Do not ask unnecessary questions. If information is missing, make strong cinematic decisions that fit the concept.
- Expand simple ideas into a compelling short-film structure (see references/director-bible.md §2).

### 2. Produce the director package
- Follow references/director-bible.md for the full methodology: story interpretation, act structure, visual bible, camera language, shot design, motion direction, transformation stages, VFX design, lighting, continuity, visual escalation, dialogue, sound, music, editing, image-to-video consistency, negative direction.
- Output the FINAL OUTPUT FORMAT (§19): TITLE, LOGLINE, VISUAL STYLE, CHARACTER BIBLE, ENVIRONMENT BIBLE, STORY, STORYBOARD (numbered shots with DURATION/ACTION/CAMERA/LIGHTING/VFX/SOUND/DIALOGUE/TRANSITION), AI VIDEO PROMPTS (one polished prompt per shot), MUSIC DIRECTION, SOUND DESIGN, FINAL HERO SHOT.
- Optionally save the package as a document with create_document (name "director_package_<title>", entity_type "documents") so it can be read back later.
- Present the package to the user and ask which shots to produce (or produce all if requested).

### 3. Generate character reference stills
- For every major character, generate one consistent reference still with generate_image (model='t2i', orientation='portrait', resolution='1080P').
- The prompt must encode the full character design from the bible: age, gender, height, body type, face structure, skin, hair, eyes, clothing, footwear, accessories, distinguishing features, default expression, plus the global style block (semi-realistic cinematic 3D adult animation, feature-film quality, dramatic lighting).
- name: "character_<slug>", entity_type: "characters".
- These stills become the identity lock for every later shot.

### 4. Generate environment reference stills
- For every major location, generate one consistent reference still with generate_image (model='t2i', orientation='landscape', resolution='1080P').
- The prompt must encode geography, architecture, time period, weather, time of day, color palette, atmosphere, and environmental storytelling details (dust in sunlight, smoke rising, distant fires, etc.).
- name: "environment_<slug>", entity_type: "locations".

### 5. Generate shot videos
- For each shot in the storyboard, generate a video clip with generate_video:
  - Default: model='ref2v' (reference mode), input_urls = the relevant character still(s) + environment still (up to 9 references), orientation='landscape', resolution='720P' (or '1080P' if the model supports it), duration = the shot's duration (2–6s).
  - The prompt must follow the AI VIDEO PROMPT structure (§6): SUBJECT, ACTION, ENVIRONMENT, CAMERA, LIGHTING, MOTION, VFX, STYLE, QUALITY — and must preserve the character/environment continuity lock.
  - If the user supplied reference images, treat them as the definitive source for identity (§17); do not redesign.
  - For explicit/adult content, use model='i2v.nsfw' instead.
  - For a single start-frame shot without multi-ref, model='i2v' is acceptable.
- name: "clip_<nn>", entity_type: "storybo