Pixar 3D Movie Maker · skill pública de Open Skill
Pixar 3D Movie Maker
Here is a detailed skill specification for **Pixar 3D Movie Maker**. This guide is formatted as a structured operational skill brief, perfect for AI video generators, virtual directors, or 3D animation production prompts. --- # Skill Specification: Pixar 3D Movie Maker ## 1. Executive Summary **Skill Name:** Pixar 3D Movie Maker **Category:** 3D Animation, Spatial Directing & AI Cinematic Production **Core Purpose:** To conceptualize, structure, and direct full-3D animated scenes and cinematic video sequences that replicate the distinct signature aesthetic, visual storytelling warmth, emotional depth, and high-fidelity CGI style of Disney-Pixar. --- ## 2. Aesthetic & Visual Style Guidelines ### A. Lighting & Color Palette * **Volumetric Warmth:** Key lighting utilizes soft, warm key lights complemented by subtle rim lights to separate characters from backgrounds. * **Color Scripting:** Saturated, harmonious color schemes that match the emotional arc of the scene (e.g., vibrant warm tones for nostalgia/comfort; rich desaturated hues for mystery or gravity). * **Subsurface Scattering (SSS):** Soft, translucent skin tones, porcelain-like surfaces, and organic texture depth across character models. ### B. Character Design & Geometry * **Stylized Realism:** Exaggerated, expressive facial features paired with realistic physics and micro-textures (e.g., woven fabric, individual hair strands, metallic gloss). * **Expressive Eye Design:** Wide, expressive eyes with dynamic specular reflections, detailed iris depth, and subtle micro-expressions. * **Soft Geometry:** Curved structural lines, slightly rounded corners, and tactile, touchable environmental surfaces. --- ## 3. Cinematography & Camera Direction * **Framing & Lensing:** Use virtual focal lengths between $35\text{mm}$ (medium cinematic shots) and $85\text{mm}$ (dramatic close-ups with shallow depth of field / creamy bokeh). * **Camera Movement:** Smooth, organic camera motions—sweeping jib moves, subtle handheld float, tracking shots at a low angle to accentuate character stature. * **Composition:** Adherence to the rule of thirds, leading lines through 3D environments, and dynamic depth layering (foreground props, midground subject, soft background scenery). --- ## 4. Operational Workflow & Scene Generation Prompt Structure When prompting or scripting scenes using this skill, structure the parameters into five core layers: | Layer | Component | Description / Parameters | | --- | --- | --- | | **01. Scene & Setup** | Environment & Props | Detailed 3D environment layout with tactile surfaces, set dressing, and micro-details. | | **02. Character Action** | Keyframe Pose & Motion | Character performance, body language, facial expression, and micro-gestures. | | **03. Visual Style** | Pixar Aesthetic | "3D Pixar animation style, RenderMan photorealistic rendering, soft volumetric lighting". | | **04. Cinematography** | Lens, Angle & Motion | Camera depth, lighting direction, focal length, track speed, and focus pull. | | **05. Atmosphere** | FX & Post-Processing | Subtle dust motes, soft ambient occlusion, cinematic bloom, and depth-of-field blur. | --- ## 5. Standardized Prompt Template ```text [Style]: Disney-Pixar 3D computer animation style, rendered in high-detail 3D CGI. [Subject]: A Pixar-style character with expressive, wide eyes, detailed micro-expressions, and soft translucent skin texturing. [Environment]: Richly detailed 3D set with soft geometry, realistic surface textures (wood grain, fabric weave, metallic sheen), and depth of field. [Lighting]: Warm volumetric key light with subtle rim lighting and soft ambient fill. [Cinematography]: 50mm cinematic lens, low-angle slow tracking shot, shallow depth of field, creamy background blur. [Quality/Engine]: RenderMan cinematic quality, soft subsurface scattering, vibrant color palette, fluid physical motion. ``` --- ## 6. Key Capabilities & Capabilities Checklist * [x] **Character Pose & Acting:** Translates emotional beats into subtle poses and facial expressions. * [x] **Lighting Consistency:** Ensures uniform illumination across character-to-environment transitions. * [x] **Texture Fidelity:** Maps realistic materials onto stylized geometry. * [x] **Storytelling Pacing:** Structures camera movements to highlight story moments smoothly.
Film
21 3 9
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4125068/2026-10-20/wgxv3cus.jpg
--- allowed-tools: - generate_image - generate_video - generate_tts - generate_music - mix_clips - split_image - recognize_image - create_document - update_document - get_document - memory_query - memory_store - memory_whoami - memory_forget - search_web --- # Pixar 3D Movie Maker ## When to use - The user wants to create a full-fledged 3D animation movie (Pixar-style) from an idea, a script, or reference images. - The user wants to advance, redo, or refine a single stage of such a movie (script, storyboard, character/location/prop, clip, voiceover, music, final cut). - Supports both portrait (9:16) short-form and landscape (16:9) cinematic; ask which if not stated. ## Reference files - `references/style-guide.md`: Pixar 3D visual style keywords + character-consistency conventions. Read before generating any image. - `references/pipeline.md`: detailed per-stage parameters and workflow. Read when running the full pipeline or a specific stage. ## Steps 1. First call `write_todos` to lay out the pipeline stages, then update the whole list as you progress (at most one in_progress). 2. Query memory with `memory_query` for any existing project background/preferences for this movie. If none, ask the user for the core idea: logline, genre/tone, main characters, target length, and orientation (portrait/landscape). 3. Run the pipeline stage by stage (see `references/pipeline.md` for per-stage detail and parameters): - **Script**: `create_document` (entity_type `scripts`) — write the full script/outline. - **Storyboard**: `generate_image` grid via `model='i2i.hd'` (entity_type `storyboard_images`), then `split_image` into panels. - **Character identity sheets**: `generate_image` (entity_type `characters`) — one 4-view sheet per main character. - **Locations**: `generate_image` (entity_type `locations`). - **Props**: `generate_image` (entity_type `props`). - **Clips**: `generate_video` with `model='ref2v'`, passing the character/location reference images as `input_urls` (entity_type `storyboard_videos`). - **Voiceover**: `generate_tts` with `model='tts.design'` + `voice_instruction` (entity_type `voiceovers`). - **Music**: `generate_music` (entity_type `music`). - **Final cut**: `mix_clips` to stitch clips and mix voiceover/music (entity_type `final_videos`). 4. Store the project background and preferences with `memory_store` (kind `project` / `preference`). 5. On completion, call `set_guidance` to suggest next actions (e.g. redo a clip, change orientation, tweak the script). ## Canvas asset conventions Use these fixed `entity_type` values per stage (never reuse across stages): - Script/outline → `scripts` - Storyboard images → `storyboard_images` - Character sheets → `characters` - Locations → `locations` - Props → `props` - Clips → `storyboard_videos` - Voiceover → `voiceovers` - Music → `music` - Final cut → `final_videos` `split_image` needs no `entity_type` (fixed to `grid_tiles`). Give each asset a short `name` (character name, shot title, track title, etc.). ## Memory - Query `memory_query` before starting to recall prior project context; don't invent details if nothing is found. - Store with `memory_store`: movie project background, genre/tone preferences, character names, orientation preference (kind `project` / `preference`). - On "forget/correct" requests: `memory_query` to find the `id`, then `memory_forget`. ## Notes - Report failures honestly; never fabricate URLs. - Video generation does not support NSFW/adult content on any channel — decline such requests politely. - Intermediate step-by-step products (storyboard grids, individual clips) will be visible on the canvas; that's expected. - Character consistency is achieved by generating identity sheets once and reusing them as `ref2v` reference images across shots.