Claude Code Plugins

Community-maintained marketplace

Feedback

veo3-image-prompt

@owainlewis/claude-skills
0
0

Analyzes images and generates optimized prompts for Google's Veo 3.1 video generation model. Use when users provide an image and want to animate it or create a video prompt from it. Applies Veo 3.1's five-part formula (Cinematography + Subject + Action + Context + Style & Ambiance) and best practices from the official prompting guide.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name veo3-image-prompt
description Analyzes images and generates optimized prompts for Google's Veo 3.1 video generation model. Use when users provide an image and want to animate it or create a video prompt from it. Applies Veo 3.1's five-part formula (Cinematography + Subject + Action + Context + Style & Ambiance) and best practices from the official prompting guide.

Veo 3.1 Image-to-Prompt Generator

Generate perfect Veo 3.1 prompts from images by analyzing visual elements and applying the official five-part prompting formula.

Core Process

When a user provides an image for Veo 3.1 prompt generation:

  1. Analyze the image thoroughly, identifying:

    • Main subjects and characters
    • Environment and setting
    • Visual style and aesthetic
    • Lighting and mood
    • Compositional elements
    • Potential actions or movements
  2. Apply the five-part formula to structure the prompt:

    • [Cinematography]: Camera work and shot composition
    • [Subject]: Main character or focal point
    • [Action]: What the subject should do
    • [Context]: Environment and background
    • [Style & Ambiance]: Aesthetic, mood, lighting
  3. Enhance with cinematic techniques:

    • Specify appropriate camera movements
    • Define shot composition (wide, close-up, etc.)
    • Include lens and focus details when relevant
    • Add audio direction (dialogue, SFX, ambient noise)
  4. Return as plain text in a code block for easy copying

Cinematography Language

Use precise cinematography terms to control tone and emotion:

Camera Movement:

  • Dolly shot, tracking shot, crane shot, aerial view
  • Slow pan, POV shot, handheld, Steadicam
  • Push in, pull out, orbit, arc shot

Composition:

  • Wide shot, medium shot, close-up, extreme close-up
  • Low angle, high angle, eye level, Dutch angle
  • Two-shot, over-the-shoulder, establishing shot

Lens & Focus:

  • Shallow depth of field, deep focus
  • Wide-angle lens, telephoto lens, macro lens
  • Soft focus, rack focus, tilt-shift

Audio Direction

Veo 3.1 generates audio based on text instructions:

Dialogue:

  • Use quotation marks: The woman says, "I've been waiting for this moment."
  • Specify tone: in a whisper, shouting excitedly, in a weary voice

Sound Effects (SFX):

  • Be specific: SFX: thunder cracks in the distance
  • Timing matters: SFX: footsteps echo on wet pavement

Ambient Noise:

  • Set the soundscape: Ambient noise: bustling city traffic and distant sirens
  • Match the environment: Ambient noise: gentle waves and seagulls

Style & Mood Guidance

Visual Styles:

  • Film-specific: shot on 35mm film, 8mm home video aesthetic, Super 8 grain
  • Era-specific: 1980s aesthetic, noir style, vintage Polaroid look
  • Genre-specific: cyberpunk, epic fantasy style, documentary style

Mood & Lighting:

  • Emotional tone: melancholic, joyful, tense, serene
  • Lighting quality: golden hour light, harsh fluorescent, soft morning light
  • Color grading: cool blue tones, warm sepia, high contrast

Output Format

Always return the prompt as plain text inside a code block for easy copying:

[Your generated Veo 3.1 prompt here]

Best Practices

DO:

  • Keep prompts concise but descriptive (2-4 sentences typical)
  • Lead with cinematography for strongest control
  • Specify what subjects ARE doing, not what they're NOT doing
  • Use quotation marks for specific dialogue
  • Include audio elements (dialogue, SFX, ambient) when appropriate
  • Match camera movement to the intended emotion

DON'T:

  • Use negative prompts (instead describe what IS present)
  • Over-specify every detail (trust the model's understanding)
  • Ignore the image's existing composition and lighting
  • Forget to specify aspect ratio if it matters (16:9 or 9:16)

Example Transformations

Image Analysis: Photo of a woman looking out a rainy window Generated Prompt:

Close-up with shallow depth of field, a young woman's face, gazing pensively out a rain-streaked window with city lights blurred in the background, inside a dimly lit apartment at night. Her reflection is faintly visible on the glass. SFX: soft rain pattering on the window. Melancholic mood with cool blue tones, cinematic.

Image Analysis: Mountain landscape at sunrise Generated Prompt:

Crane shot starting low on a lone hiker standing at the mountain summit and ascending high above, revealing a vast range of misty peaks bathed in golden sunrise light. The hiker raises their arms in triumph. SFX: wind whistling across the peaks. Epic, inspirational mood with warm morning light.

Advanced Features

When users need more complex control, reference these Veo 3.1 capabilities:

  • First and Last Frame: Create transitions between two images
  • Ingredients to Video: Maintain consistency with reference images
  • Timestamp Prompting: Multi-shot sequences (e.g., [00:00-00:02] ...)
  • Add/Remove Object: Modify generated videos (note: uses Veo 2, no audio)

For detailed guidance on these workflows, see references/advanced-workflows.md.