AI Artist - Prompt Engineering + Image Generation
Comprehensive prompt engineering with integrated image generation. Contains 6,000+ curated prompts, dual-option generation workflow, and searchable database with BM25 ranking.
Dual-Option Workflow
When generating images, always offer both options:
| Option |
Model |
Best For |
Speed |
| Standard |
gemini-2.5-flash-image |
Search-based prompts, known formats |
Fast |
| Creative |
gemini-3-pro-image-preview |
Unique art direction, complex scenes |
Best quality |
Quick Start
Option 1: Standard (Search + Flash)
# 1. Search for similar prompts
python3 .claude/skills/ai-artist/scripts/search.py "<keywords>" --domain examples
# 2. Generate with Flash model
.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
--task generate --model gemini-2.5-flash-image \
--prompt "<adapted_prompt>" --aspect-ratio <ratio> --size 2K --output <path>
Option 2: Creative (Art Direction + Pro)
# Generate with Pro model (4K, complex prompts)
.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
--task generate --model gemini-3-pro-image-preview \
--prompt "<creative_prompt>" --aspect-ratio <ratio> --size 4K --output <path>
Arguments
| Arg |
Values |
Default |
--model |
gemini-2.5-flash-image (Standard), gemini-3-pro-image-preview (Creative) |
Flash |
--aspect-ratio |
1:1, 16:9, 9:16, 3:4, 4:3, 2:3, 3:2 |
1:1 |
--size |
1K, 2K, 4K (4K only Pro) |
2K |
--num-images |
1-4 |
1 |
--output |
file path |
auto |
Search Domains
python3 .claude/skills/ai-artist/scripts/search.py "<query>" --domain <domain>
| Domain |
Use For |
examples |
6000+ curated prompts (quote cards, bento grids, thumbnails) |
style |
Visual aesthetics (cyberpunk, minimalist, cinematic) |
subject |
Subject tips (portrait, product, landscape) |
platform |
Platform syntax (midjourney, dall-e, sd) |
Creative Art Directions
For Creative option, apply unique styles:
| Style |
Key Elements |
| Cyberpunk Neon |
Holographic, neon glow, rain, cityscape, Japanese text |
| Vaporwave |
Greek statues, sunset grid, Windows 95, chrome, dolphins |
| Isometric 3D |
Tilt-shift, miniature diorama, cute buildings, floating island |
| Bento Grid |
Apple liquid glass, transparent cards, gradient glow |
| Editorial |
Clean white, bold typography, magazine layout |
Example Workflow
User: "Create marketing infographic"
Standard Option:
python3 .claude/skills/ai-artist/scripts/search.py "bento grid marketing" --domain examples
# → Found: "Premium liquid glass Bento grid product infographic"
# → Adapt prompt, generate with Flash
Creative Option:
# Cyberpunk art direction with Pro model
.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
--task generate --model gemini-3-pro-image-preview \
--prompt "Create a cyberpunk neon holographic infographic titled 'MARKETING 2026'. Style: Blade Runner aesthetic, dark cityscape, rain, neon data cards floating around holographic AI head. Hyper-detailed, cinematic." \
--aspect-ratio 1:1 --size 4K --output infographic-cyberpunk.png
References
| Topic |
File |
| Image generation workflow |
references/image-generation-workflow.md |
| Nano Banana models |
references/nano-banana.md |
| Prompt examples (6000+) |
references/nano-banana-pro-examples.md |
| LLM patterns |
references/llm-prompting.md |
| Image syntax |
references/image-prompting.md |
| Domain patterns |
references/domain-patterns.md |
Anti-Patterns
| Issue |
Solution |
| Only offering one option |
Always present Standard AND Creative |
| Using Flash for complex art |
Use Pro for unique art direction |
| Vague creative prompts |
Add detailed style, layout, elements |
| Missing output path |
Always specify --output |
1---2name: ai-artist-23description: Prompt + image gen. 6000+ examples. Dual-option: Standard (Flash) or Creative (Pro). Generate infographics, thumbnails, avatars, product shots with Nano Banana.4license: MIT5---6
7# AI Artist - Prompt Engineering + Image Generation
8
9Comprehensive prompt engineering with integrated image generation. Contains 6,000+ curated prompts, dual-option generation workflow, and searchable database with BM25 ranking.
10
11## Dual-Option Workflow
12
13When generating images, **always offer both options**:
14
15| Option | Model | Best For | Speed |
16|--------|-------|----------|-------|
17| **Standard** | `gemini-2.5-flash-image` | Search-based prompts, known formats | Fast |
18| **Creative** | `gemini-3-pro-image-preview` | Unique art direction, complex scenes | Best quality |
19
20---
21
22## Quick Start
23
24### Option 1: Standard (Search + Flash)
25
26```bash
27# 1. Search for similar prompts
28python3 .claude/skills/ai-artist/scripts/search.py "<keywords>" --domain examples
29
30# 2. Generate with Flash model
31.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
32 --task generate --model gemini-2.5-flash-image \
33 --prompt "<adapted_prompt>" --aspect-ratio <ratio> --size 2K --output <path>
34```
35
36### Option 2: Creative (Art Direction + Pro)
37
38```bash
39# Generate with Pro model (4K, complex prompts)
40.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
41 --task generate --model gemini-3-pro-image-preview \
42 --prompt "<creative_prompt>" --aspect-ratio <ratio> --size 4K --output <path>
43```
44
45---
46
47## Arguments
48
49| Arg | Values | Default |
50|-----|--------|---------|
51| `--model` | `gemini-2.5-flash-image` (Standard), `gemini-3-pro-image-preview` (Creative) | Flash |
52| `--aspect-ratio` | `1:1`, `16:9`, `9:16`, `3:4`, `4:3`, `2:3`, `3:2` | `1:1` |
53| `--size` | `1K`, `2K`, `4K` (4K only Pro) | `2K` |
54| `--num-images` | `1-4` | `1` |
55| `--output` | file path | auto |
56
57---
58
59## Search Domains
60
61```bash
62python3 .claude/skills/ai-artist/scripts/search.py "<query>" --domain <domain>
63```
64
65| Domain | Use For |
66|--------|---------|
67| `examples` | 6000+ curated prompts (quote cards, bento grids, thumbnails) |
68| `style` | Visual aesthetics (cyberpunk, minimalist, cinematic) |
69| `subject` | Subject tips (portrait, product, landscape) |
70| `platform` | Platform syntax (midjourney, dall-e, sd) |
71
72---
73
74## Creative Art Directions
75
76For Creative option, apply unique styles:
77
78| Style | Key Elements |
79|-------|--------------|
80| **Cyberpunk Neon** | Holographic, neon glow, rain, cityscape, Japanese text |
81| **Vaporwave** | Greek statues, sunset grid, Windows 95, chrome, dolphins |
82| **Isometric 3D** | Tilt-shift, miniature diorama, cute buildings, floating island |
83| **Bento Grid** | Apple liquid glass, transparent cards, gradient glow |
84| **Editorial** | Clean white, bold typography, magazine layout |
85
86---
87
88## Example Workflow
89
90### User: "Create marketing infographic"
91
92**Standard Option:**
93```bash
94python3 .claude/skills/ai-artist/scripts/search.py "bento grid marketing" --domain examples
95# → Found: "Premium liquid glass Bento grid product infographic"
96# → Adapt prompt, generate with Flash
97```
98
99**Creative Option:**
100```bash
101# Cyberpunk art direction with Pro model
102.claude/skills/.venv/bin/python3 .claude/skills/ai-multimodal/scripts/gemini_batch_process.py \
103 --task generate --model gemini-3-pro-image-preview \
104 --prompt "Create a cyberpunk neon holographic infographic titled 'MARKETING 2026'. Style: Blade Runner aesthetic, dark cityscape, rain, neon data cards floating around holographic AI head. Hyper-detailed, cinematic." \
105 --aspect-ratio 1:1 --size 4K --output infographic-cyberpunk.png
106```
107
108---
109
110## References
111
112| Topic | File |
113|-------|------|
114| Image generation workflow | `references/image-generation-workflow.md` |
115| Nano Banana models | `references/nano-banana.md` |
116| Prompt examples (6000+) | `references/nano-banana-pro-examples.md` |
117| LLM patterns | `references/llm-prompting.md` |
118| Image syntax | `references/image-prompting.md` |
119| Domain patterns | `references/domain-patterns.md` |
120
121---
122
123## Anti-Patterns
124
125| Issue | Solution |
126|-------|----------|
127| Only offering one option | Always present Standard AND Creative |
128| Using Flash for complex art | Use Pro for unique art direction |
129| Vague creative prompts | Add detailed style, layout, elements |
130| Missing output path | Always specify `--output` |
majiayu000