Image Prompt: Agent Skill (SKILL.md) for Claude Code, Codex and Cursor

Writes image-generation prompts (subject, setting, light, lens, style, aspect ratio) for product shots, portraits and ads.

Prompt

Installs to .agents/skills/image-prompt-agent-skill-skill-md-for-claude-code-codex-and-curs/

mkdir -p .agents/skills/image-prompt-agent-skill-skill-md-for-claude-code-codex-and-curs && curl -sL "https://prompts-gpt.com/prompts/image-prompt-agent-skill-skill-md-for-claude-code-codex-and-cursor/raw" -o .agents/skills/image-prompt-agent-skill-skill-md-for-claude-code-codex-and-curs/SKILL.md
SKILL.md
Preview
---
name: image-prompt-agent-skill-skill-md-for-claude-code-codex-and-curs
description: "Writes image-generation prompts (subject, setting, light, lens, style, aspect ratio) for product shots, portraits and ads."
---

# Image Prompt: Agent Skill (SKILL.md) for Claude Code, Codex and Cursor

Writes image-generation prompts (subject, setting, light, lens, style, aspect ratio) for product shots, portraits and ads.

## Prompt

```
---
name: prompts-gpt-image-prompt
description: Writes image-generation prompts (subject, setting, light, lens, style, aspect ratio) for product shots, portraits and ads. Use when asked for an image prompt or to describe a photo for an image model.
---

# Image prompt

## Steps

1. Get the brief: what the image is for (ad, thumbnail, product page, portrait), the subject, the target model (ChatGPT image, Gemini / Nano Banana, Midjourney, Flux) and the aspect ratio. If the subject or purpose is missing, ask one short question; otherwise assume and say what you assumed.
2. Describe the subject precisely: materials, colours, shape, scale, angle. For an edit of an uploaded photo, say what must stay exactly the same (face, logo, product shape) before what changes.
3. Set the scene and light: location, time of day, light source and softness, background.
4. Pick the camera: lens length, depth of field, framing and camera height.
5. Add style and mood in at most five words. For text in the image, give the exact words in quotes and where they go.
6. Adapt to the model: Midjourney takes parameters at the end (for example --ar 4:5); ChatGPT and Gemini read plain sentences.

## Output format

A numbered list with one prompt per requested style or variant, each under 80 words, then one line per prompt saying what to change if the result misses (light, angle or crop).

## Rules

- Concrete nouns over adjectives. No artist names or living photographers' styles.
- Never ask for a real person's likeness, a brand's logo or trademarked characters unless the user owns them.
- State any assumption in one line at the end.
```

## Usage Notes

Save as .claude/skills/prompts-gpt-image-prompt/SKILL.md in one project (or ~/.claude/skills/prompts-gpt-image-prompt/ for all of them); Codex reads .agents/skills/prompts-gpt-image-prompt/SKILL.md. Then ask for the job in plain words and the agent loads it. Or install with: npx skills add prompts-gpt/skills --skill prompts-gpt-image-prompt.

Example output

At a glanceThree image prompts for the Halde dripperEach prompt fixes setting, light, lens and style for a 4:5 frame
ItemProduct hero (Midjourney)Headline ad (ChatGPT image)
SettingPale oak counter, off-white plaster wallSage linen cloth, terracotta backdrop
LightSoft morning window light from leftEven overhead diffuser light
Lens90mm, f/4, eye level70mm, f/8, three-quarter view
TextNone, negative space above"Brew Slower." in top third
MoodCalm, minimal, ScandinavianBold, graphic, editorial

Example for: a product-page hero and an Instagram ad for "Halde," a fictional matte-ceramic pour-over coffee dripper in sage green; target models Midjourney and ChatGPT image; 4:5 aspect ratio.

  1. Product page hero (Midjourney)
    A matte sage-green ceramic pour-over dripper, conical with three thin vertical ribs, about 12 cm tall, set on a clear glass carafe half filled with dark coffee. A paper filter is folded inside. The scene is a pale oak countertop against a warm off-white plaster wall. Soft morning window light comes from the left with a gentle shadow to the right. Shot at 90mm, f/4, eye level, centered, with negative space above. Calm, minimal, Scandinavian kitchen. --ar 4:5 --style raw
  2. Lifestyle ad (ChatGPT image)
    A close-up of hands pouring hot water from a black gooseneck kettle into a matte sage-green ceramic dripper resting on a glass carafe. Steam rises in a thin curl. The setting is a sunlit kitchen windowsill with a small potted rosemary, blurred in the background. Low golden-hour light comes from behind and rims the steam. The camera is a 50mm lens at f/2, slightly above the dripper, with a shallow focus on the spout and water stream. Warm, slow, quiet morning ritual.
  3. Ad with headline text (ChatGPT image)
    A matte sage-green ceramic pour-over dripper on a glass carafe, shown in a three-quarter view from the left. It sits on a sage linen cloth against a flat terracotta backdrop. Soft, even studio light comes from a large overhead diffuser, with a faint contact shadow underneath. The lens is 70mm at f/8 and frames the whole product sharply in the lower two-thirds. The exact text "Brew Slower." appears in clean white sans-serif, centered in the top third. Bold, graphic, editorial.

If the result misses:

  • Prompt 1: If the light looks flat, change it to "hard side light at a low angle, a defined shadow falling right."
  • Prompt 2: If the hands dominate the frame, change it to "crop tighter on the spout and dripper, with hands only partly in frame."
  • Prompt 3: If the text overlaps the product, change it to "product in the bottom half only, with a wide empty band at the top for text."

Assumption: the dripper is the user's own unbranded product, so no logo is shown.

Written by AI from this prompt. Your results will vary.

Save as .claude/skills/prompts-gpt-image-prompt/SKILL.md in one project (or ~/.claude/skills/prompts-gpt-image-prompt/ for all of them); Codex reads .agents/skills/prompts-gpt-image-prompt/SKILL.md. Then ask for the job in plain words and the agent loads it. Or install with: npx skills add prompts-gpt/skills --skill prompts-gpt-image-prompt.

Make it repeatable

Keep using this prompt without pasting it in every time.

Prompt details

By
Prompts-GPT editors
Published
How prompts are written and checked

Did this prompt work?

Keep exploring

More in AI Marketing