Muapi Cinema Director

🎬 AI Cinema Director Skill

A specialized skill for AI Agents to direct high-fidelity cinematic video. The Cinema Director skill translates high-level creative intent into technical cinematographic directives for state-of-the-art video models (Veo3, Kling, Luma).

Core Competencies

1. Shot Composition Analysis: Mapping emotional beats to appropriate framing (e.g., Extreme Close-Up for intimacy, Wide Shot for isolation).
2. Camera Movement Orchestration: Directing complex physical movements (Dolly, Truck, Crane) and lens-based effects (Rack Focus, Dolly Zoom).
3. Lighting & Atmosphere Design: Specifying temporal and stylistic lighting (Golden Hour, Chiaroscuro, Volumetric God Rays).
4. Technical Parameter Optimization: Automatically selecting optimal frame rates, aspect ratios, and model-specific biases.

---

🏗️ Technical Specification

1. Intent Mapping Table

| Creative Intent | Framing | Movement | Lighting | | :--- | :--- | :--- | :--- | | Heroic Reveal | Low Angle / Wide | Crane Up / Orbit | Rim Lighting / High Contrast | | Tense/Uneasy | Dutch Angle | Handheld Shake | Low Key / Harsh Shadows | | Introspective | Close-Up | Slow Push In | Soft Rembrandt / Window Light | | Majestic/Epic | Extreme Wide | Drone Flyover | Golden Hour / Volumetric | | Melancholic | Profile / Medium | Slow Pull Out | Blue Hour / Desaturated |

2. Physical Camera Movements

• Dolly In/Out: Physical camera movement on a track toward/away from the subject.
• Truck Left/Right: Lateral physical movement.
• Crane/Jib: Sweeping vertical movement from a height.
• Orbit: Circular movement around a center point.
• Pedestal: Vertical elevation change (without tilting).

3. Lens & Optical Controls

• Shallow DOF: Background blur (Bokeh).
• Anamorphic: Horizontal flares and wide-screen cinematic feel.
• Rack Focus: Shifting focus between planes within the shot.

---

🧠 Prompt Optimization Protocol (Agent Instruction)

Before calling the script, the Agent MUST expand the user's raw prompt using these Director's Rules:

1. TECHNICAL INFUSION: Transform "cool action" into a technical shot.

• Formula: [Shot Type] + [Subject/Action] + [Environment] + [Lighting] + [Camera Movement] + [Lens Effect]

1. MOTION DYNAMICS: Use cinematic verbs: Dolly In (intimacy), Crane Up (majestic), Orbit (heroic), Truck (parallel motion).
2. LIGHTING RECIPES: Apply specific illumination: Volumetric God Rays, Teal-and-Orange Grade, Cyberpunk Rim Lighting, Rembrandt Portrait Lighting.
3. PHYSICS LOGIC: Describe light relationships (e.g., "Neon reflections shimmering on rain-slicked asphalt") to trigger model reasoning.

---

🚀 Protocol: Using the Cinema Director

Step 1: Define the Creative Brief

Provide the agent with a subject and a "Director's Intent."

Step 2: Invoke the Script

The generate-film.sh script accepts a --brief which it expands using its internal knowledge of cinematography.

# Directing a scene
bash scripts/generate-film.sh 
  --subject "A lone samurai in a blizzard" 
  --intent "epic reveal" 
  --model "kling-master"

Step 3: Handle the Async Response

Video generation is asynchronous. Use the returned request_id to poll for completion via core/platform/check-result.sh.

---

⚠️ Constraints & Guardrails

• Temporal Consistency: Avoid complex subject transformations in a single shot (e.g., "man turns into a bird").
• Movement Collisions: Do not combine contradictory movements (e.g., "Dolly In" and "Dolly Out" simultaneously).
• Physical Realism: Prefer movements possible with real-world equipment for a more professional "film" look.
• Model Bias:
• Veo3: Best for slow, high-quality aesthetic shots.
• Kling: Best for complex character motion and physics.
• Luma: Best for fast-paced, high-action cinematic sequences.

---

⚙️ Implementation Details

This skill acts as an "Expert Translator" for the core/media/generate-video.sh primitive. It maintains a dictionary of cinematic styles and injects technical directives into the prompt before execution.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: SamurAIGPT
• Source: SamurAIGPT/Generative-Media-Skills
• License: MIT
• Homepage: https://muapi.ai?utmsource=github&utmmedium=about&utm_campaign=generative-media-skills

Install and usage instructions live in the source repository linked above.