name	image-prompt-cinematography
description	Cinematic image and video prompt discipline
version	1.0.0
metadata	{"hermes":{"category":"creative","tags":["image-generation","video-generation","cinematography","prompting"]}}

Image Prompt Cinematography

Turn visual ideas into disciplined, character-first prompt language — for image generation, video generation, character sheets, shot planning, and frame diagnostics.

The mantra: the world is maximal, the frame is merciless.

When to Use

The user asks for an image or video prompt, a visual spec, or direct generation.
The user asks for character design sheets, shot language, or color direction.
A generated image came back wrong and needs diagnosis before re-rolling.
Any visual work where the default output would be AI-flavored clutter.

Procedure

Start with the character. Approved character specs override style flourish. If no spec exists, establish silhouette and must-have details first.
One emotional action only. Not a mood board — a single readable moment with physical logic.
Specify composition explicitly. Framing, camera distance, angle, lens feeling, screen-left/screen-right logic, which hand does what, depth of field.
Give the environment one temperament — and make it argue with the character. Friction is the drama.
Choose the chromatic argument. Warm vs. cool dominance, a color intrusion, desaturation, or the one wrong color. Color is an argument, not decoration.
Apply the visual tax. State what the moment costs the body physically. Strength without cost reads as plastic.
Remove everything that competes with the subject. Props and background incidents go before character clarity does.
State what the frame refuses to show. Withholding is power.
Compile in this order: Character → Action → Composition → Environment → Lighting → Style → Withholding. (See references/prompt-template.md.)
If asked to generate, generate — don't return prompt text when the user wanted the image.

Pitfalls

AI-default clutter: decorative worldbuilding noise, three light sources, props that earn nothing. Cut before compiling.
Effortless power posing: bodies that pay no tax for what they do. Always state the cost.
Everything has temperament: objects and environments get emotional adjectives, or they read as catalog photography.
Two emotional engines: a frame arguing with itself. One dominant subject, one readable silhouette, one engine.
Style-first prompting: style flourish before character accuracy produces pretty strangers.
Bloat spiral: if the prompt grows past compact, remove environment incidents first, then props — never character clarity.

Verification

Before delivering a compiled prompt, confirm:

One dominant subject with a readable silhouette
One emotional action with visible physical logic
Explicit composition: distance, angle, screen direction, hand logic
Environment has exactly one temperament and it argues with the character
The chromatic argument is named
The visual tax is stated
The withholding is stated — the frame refuses something
Nothing in the prompt competes with the subject

Image Prompt Cinematography

Turn visual ideas into disciplined, character-first prompt language — for image generation, video generation, character sheets, shot planning, and frame diagnostics.

The mantra: the world is maximal, the frame is merciless.

When to Use

The user asks for an image or video prompt, a visual spec, or direct generation.

The user asks for character design sheets, shot language, or color direction.

A generated image came back wrong and needs diagnosis before re-rolling.

Any visual work where the default output would be AI-flavored clutter.

Procedure

Start with the character. Approved character specs override style flourish. If no spec exists, establish silhouette and must-have details first.

One emotional action only. Not a mood board — a single readable moment with physical logic.

Specify composition explicitly. Framing, camera distance, angle, lens feeling, screen-left/screen-right logic, which hand does what, depth of field.

Give the environment one temperament — and make it argue with the character. Friction is the drama.

Choose the chromatic argument. Warm vs. cool dominance, a color intrusion, desaturation, or the one wrong color. Color is an argument, not decoration.

Apply the visual tax. State what the moment costs the body physically. Strength without cost reads as plastic.

Remove everything that competes with the subject. Props and background incidents go before character clarity does.

State what the frame refuses to show. Withholding is power.

Compile in this order: Character → Action → Composition → Environment → Lighting → Style → Withholding. (See references/prompt-template.md.)

If asked to generate, generate — don't return prompt text when the user wanted the image.

Pitfalls

AI-default clutter: decorative worldbuilding noise, three light sources, props that earn nothing. Cut before compiling.

Effortless power posing: bodies that pay no tax for what they do. Always state the cost.

Everything has temperament: objects and environments get emotional adjectives, or they read as catalog photography.

Two emotional engines: a frame arguing with itself. One dominant subject, one readable silhouette, one engine.

Style-first prompting: style flourish before character accuracy produces pretty strangers.

Bloat spiral: if the prompt grows past compact, remove environment incidents first, then props — never character clarity.

Verification

Before delivering a compiled prompt, confirm:

One dominant subject with a readable silhouette

One emotional action with visible physical logic

Explicit composition: distance, angle, screen direction, hand logic

Environment has exactly one temperament and it argues with the character

The chromatic argument is named

The visual tax is stated

The withholding is stated — the frame refuses something

Nothing in the prompt competes with the subject

image-prompt-cinematography