Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →How to combine two images or ideas into one, using multi-image reference input, the modern replacement for photobashing.
.claude/skills/whatsuppiyush-ai-blending-reference-images/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 133% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -2% | 0% |
Combining two images or ideas, a face and a galaxy, a product and a scene, two photos into one, used to mean manual photobashing or fiddly Midjourney blends. Now the reliable route is multi-image reference input: give the model the images and describe how to combine them. This guide covers the methods and the prompt patterns that make blends look intentional rather than pasted.
Works in: Midjourney, GPT Image, Flux, Nano Banana, Stable Diffusion.
When the user asks for an image prompt in this style, compose one using the formulas and vocabulary below. Fill each slot with a concrete choice, then return the finished prompt. Prefer naming a real camera/lens, a light, and a colour or film treatment, which is what makes the output read as a real photograph rather than an AI render.
The blend recipe
[image A] + [image B] + how they combine (double exposure / composited into / textured with)The relationship word matters: 'a double exposure of X and Y' reads differently from 'X composited into Y' or 'X with the texture of Y'. Name the one you want.
_Reference blend vs word blend_
Reference (GPT Image / Nano Banana): [attach face image] + [attach forest image], blend them as a double exposure, face filled with the forest. · Words only: 'double exposure portrait of a woman's profile filled with a misty pine forest, elegant, monochrome'._Double exposure_
Double exposure portrait of a woman's profile filled with a misty pine forest, elegant, monochrome, blended seamlessly_Texture blend_
A portrait of a woman blended with autumn leaves, her skin taking on the texture and colour of the leaves, artistic, seamless_Composite into a scene_
A perfume bottle composited into a dramatic mountain landscape at sunset, the product integrated with matching light and reflections, seamlessMulti-image reference (most control)
Double exposure (words only)
Compositing a product into a scene
From God of Skills: a curated, hand-tested directory of AI skills, prompts, templates and image style guides. Source: https://godofskills.com/guides/ai-blending-reference-images?ref=claude-skill
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 12,273 | 7,583 | -38% | 1 | 1 | 0% | 2,008 | 1,999 | -0% | 0 | 0 | — |
case-02 | fail→fail | 5,687 | 5,469 | -4% | 1 | 1 | 0% | 980 | 1,685 | +72% | 0 | 0 | — |
case-03 | fail→fail | 9,815 | 4,895 | -50% | 1 | 1 | 0% | 1,552 | 1,560 | +1% | 0 | 0 | — |
case-04 | fail→fail | 8,960 | 5,569 | -38% | 1 | 1 | 0% | 1,591 | 1,639 | +3% | 0 | 0 | — |
case-05 | fail→fail | 10,427 | 5,416 | -48% | 1 | 1 | 0% | 1,721 | 1,542 | -10% | 0 | 0 | — |
case-06 | pass→pass | 10,267 | 5,995 | -42% | 1 | 1 | 0% | 1,757 | 1,856 | +6% | 0 | 0 | — |
case-07 | pass→pass | 8,277 | 6,885 | -17% | 1 | 1 | 0% | 1,417 | 1,939 | +37% | 0 | 0 | — |
case-08 | fail→fail | 10,126 | 7,252 | -28% | 1 | 1 | 0% | 1,675 | 2,001 | +19% | 0 | 0 | — |
case-09 | fail→pass | 3,460 | 3,361 | -3% | 1 | 1 | 0% | 530 | 1,234 | +133% | 0 | 0 | — |
case-10 | fail→fail | 8,594 | 4,164 | -52% | 1 | 1 | 0% | 1,397 | 1,455 | +4% | 0 | 0 | — |
case-11 | fail→fail | 9,109 | 6,336 | -30% | 1 | 1 | 0% | 1,400 | 1,792 | +28% | 0 | 0 | — |
case-12 | fail→fail | 6,873 | 5,097 | -26% | 1 | 1 | 0% | 1,112 | 1,637 | +47% | 0 | 0 | — |
case-13 | fail→fail | 8,079 | 4,043 | -50% | 1 | 1 | 0% | 1,462 | 1,437 | -2% | 0 | 0 | — |
case-14 | fail→pass | 9,024 | 7,692 | -15% | 1 | 1 | 0% | 1,671 | 2,078 | +24% | 0 | 0 | — |
case-15 | fail→pass | 13,042 | 7,219 | -45% | 1 | 1 | 0% | 1,929 | 1,919 | -1% | 0 | 0 | — |
case-16 | fail→fail | 8,533 | 5,882 | -31% | 1 | 1 | 0% | 1,460 | 1,602 | +10% | 0 | 0 | — |
case-17 | fail→pass | 13,345 | 3,028 | -77% | 1 | 1 | 0% | 2,218 | 1,250 | -44% | 0 | 0 | — |
case-18 | fail→fail | 8,125 | 5,886 | -28% | 1 | 1 | 0% | 1,308 | 1,522 | +16% | 0 | 0 | — |
case-19 | fail→fail | 8,598 | 5,584 | -35% | 1 | 1 | 0% | 1,559 | 1,726 | +11% | 0 | 0 | — |
case-20 | fail→fail | 4,852 | 4,173 | -14% | 1 | 1 | 0% | 851 | 1,391 | +63% | 0 | 0 | — |
case-21 | fail→pass | 11,817 | 5,943 | -50% | 1 | 1 | 0% | 1,780 | 1,742 | -2% | 0 | 0 | — |
case-22 | pass→pass | 17,550 | 11,689 | -33% | 1 | 1 | 0% | 2,975 | 2,669 | -10% | 0 | 0 | — |
case-23 | pass→pass | 19,056 | 14,892 | -22% | 1 | 1 | 0% | 3,009 | 3,348 | +11% | 0 | 0 | — |
case-24 | pass→pass | 9,862 | 6,628 | -33% | 1 | 1 | 0% | 1,443 | 1,787 | +24% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +21 percentage points is the difference between those two pass rates over the 24 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/9/2026 | +40% |
| gemini-3.6-flash | verified | 8/6/2026 | +43% |
Other measured skills in the registry, with their headline benchmark lift.