docs: Add photorealistic aging portrait example

Shows multi-panel consistency for photorealistic people:
- JSON structure with face preservation
- Same person at ages 10, 20, 40, 80
- Heterochromia + freckles maintained across panels
- Real generated result

Also added to SKILL.md intro examples.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Matt Van Horn
2026-01-24 16:28:15 -08:00
parent 7f1ab7a2e9
commit 6831e624b8
3 changed files with 64 additions and 1 deletions
+63
View File
@@ -277,6 +277,69 @@ This prompt was generated based on research from 10 Reddit threads (294 upvotes)
--- ---
## Example: Photorealistic Portraits (Aging Grid)
**Query:** `/last30days photorealistic people in nano banana pro`
**Research Output:**
> JSON prompts are the dominant format for photorealistic people in Nano Banana Pro. Nearly every viral X post (4,400+ likes total) uses structured JSON with specific fields rather than natural language. The key is specifying demographics, skin texture, lighting, and camera settings in a structured way.
>
> Recent caveat: Reddit reports Nano Banana Pro was "nerfed" in early January 2026 with reduced resolution and stricter moderation on faces. Starting a new chat can help if quality drops mid-session.
**Key patterns discovered:**
1. JSON structure with prompt_type, subject, lighting, camera_settings fields
2. Skin texture keywords: "natural pores", "realistic skin texture", "8K texture quality"
3. Face preservation: "preserve_original": true when using reference photos
4. Camera realism: specify lens (85mm f/1.4), depth of field, lighting setup
5. Wet skin close-ups are trending for maximum realism
6. Negative prompts: "no cartoon effect, no distortion, correct anatomy"
**Then asked:** "2x2 grid of the same person at age 10, 20, 40 and 80 - a woman with brown hair and a lot of freckles and cat eyes that are a different color"
**Generated Prompt:**
```json
{
"prompt_type": "Ultra-Photorealistic Multi-Panel Portrait",
"layout": "2x2 grid, seamless white dividers",
"aspect_ratio": "1:1",
"consistency": {
"same_person": true,
"preserve_features": ["bone structure", "freckle pattern", "eye shape", "heterochromia"]
},
"subject": {
"identity": "Same woman shown at 4 life stages",
"distinctive_features": {
"eyes": "Almond-shaped cat eyes with heterochromia - one green, one amber",
"skin": "Fair complexion densely covered in natural freckles across nose, cheeks, and forehead",
"hair": "Brown hair, natural texture"
}
},
"panels": [
{"position": "top-left", "age": 10, "expression": "curious childhood smile", "hair": "long brown hair with bangs"},
{"position": "top-right", "age": 20, "expression": "confident young adult", "hair": "shoulder-length brown hair"},
{"position": "bottom-left", "age": 40, "expression": "warm maternal gaze", "hair": "brown hair with subtle highlights"},
{"position": "bottom-right", "age": 80, "expression": "wise gentle smile", "hair": "silver-gray, formerly brown"}
],
"lighting": "Soft natural window light, consistent across all panels",
"camera": "85mm portrait lens, shallow depth of field, neutral background",
"texture_quality": "8K, natural skin texture with visible pores, realistic freckle detail",
"negative": "no cartoon effect, no airbrushed skin, no symmetrical freckles, maintain heterochromia in all panels"
}
```
**The Result:**
<p align="center">
<img src="assets/aging-portrait.jpeg" width="500" alt="Aging portrait grid - same woman at 10, 20, 40, and 80"/>
</p>
Same freckles. Same heterochromia. Same bone structure. Four life stages in one coherent image.
This example shows /last30days learning **advanced image generation patterns** (multi-panel consistency, face preservation across variations) and applying them to create a complex composite image.
---
## Example: Best Rap Songs Right Now ## Example: Best Rap Songs Right Now
**Query:** `/last30days What are the best rap songs lately` **Query:** `/last30days What are the best rap songs lately`
+1 -1
View File
@@ -13,7 +13,7 @@ allowed-tools: Bash, Read, Write, AskUserQuestion, WebSearch
Research ANY topic across Reddit, X, and the web. Surface what people are actually discussing, recommending, and debating right now. Research ANY topic across Reddit, X, and the web. Surface what people are actually discussing, recommending, and debating right now.
Use cases: Use cases:
- **Prompting**: "Midjourney prompts", "ChatGPT image generation" → learn techniques, get copy-paste prompts - **Prompting**: "photorealistic people in Nano Banana Pro", "Midjourney prompts", "ChatGPT image generation" → learn techniques, get copy-paste prompts
- **Recommendations**: "best Claude Code skills", "top AI tools" → get a LIST of specific things people mention - **Recommendations**: "best Claude Code skills", "top AI tools" → get a LIST of specific things people mention
- **News**: "what's happening with OpenAI", "latest AI announcements" → current events and updates - **News**: "what's happening with OpenAI", "latest AI announcements" → current events and updates
- **General**: any topic you're curious about → understand what the community is saying - **General**: any topic you're curious about → understand what the community is saying
Binary file not shown.

After

Width:  |  Height:  |  Size: 2.7 MiB