nano-banana-pro

name	nano-banana-pro
description	Nano Banana 2 / Gemini 3.1 Flash Image gen/edit: text-to-image, image-to-image, 512/1K/2K/4K, input-image workflows.

Source: https://github.com/steipete/agent-scripts/tree/main/skills/nano-banana-pro (adopted 2026-05-16) 적용 대상: Gemini 3.1 Flash Image (Nano Banana 2) API 호출. 기존 image-gen 스킬은 Codex OAuth (OpenAI 이미지). 본 스킬은 Gemini API key 기반 — 다른 백엔드.

Nano Banana 2 Image Generation & Editing

Generate new images or edit existing ones using Google's Nano Banana 2 API (Gemini 3.1 Flash Image).

Script path (adapted)

~/2lab.ai/skills/nano-banana-pro/scripts/generate_image.py

Usage

Run the script using absolute path (do NOT cd to skill directory first):

Generate new image:

uv run ~/2lab.ai/skills/nano-banana-pro/scripts/generate_image.py --prompt "your image description" --filename "output-name.png" [--resolution 512|1K|2K|4K] [--api-key KEY]

Edit existing image:

uv run ~/2lab.ai/skills/nano-banana-pro/scripts/generate_image.py --prompt "editing instructions" --filename "output-name.png" --input-image "path/to/input.png" [--resolution 512|1K|2K|4K] [--api-key KEY]

Important: Always run from 지혁's current working directory so images are saved where 지혁 is working, not in the skill directory. Default 저장 위치: ARTIFACTS/images/YYYY-MM/.

Default Workflow (draft → iterate → final)

Goal: fast iteration without burning time on 4K until the prompt is correct.

• Draft (1K): quick feedback loop
• Iterate: adjust prompt in small diffs; keep filename new per run
  • If editing: keep the same --input-image for every iteration until 지혁 is happy.
• Final (4K): only when prompt is locked

Resolution Options

The Gemini 3.1 Flash Image API supports these output size values:

• 512 - compact ~512px-class resolution
• 1K (default) - ~1024px resolution
• 2K - ~2048px resolution
• 4K - ~4096px resolution

Map user requests to API parameters:

• "512", "512px", "0.5K", "thumbnail", "tiny" → 512
• No mention of resolution → 1K
• "low resolution", "1080", "1080p", "1K" → 1K
• "2K", "2048", "normal", "medium resolution" → 2K
• "high resolution", "high-res", "hi-res", "4K", "ultra" → 4K

API Key

The script checks for API key in this order:

1. --api-key argument (use if 지혁 provided key in chat)
2. GEMINI_API_KEY environment variable

If neither is available, the script exits with an error message.

지혁 환경에서는 .env 또는 ~/2lab.ai/p9/.env에 GEMINI_API_KEY 둘 것 (이미 있을 수 있음).

Preflight + Common Failures (fast fixes)

• Preflight:

  • command -v uv (must exist)
  • test -n "$GEMINI_API_KEY" (or pass --api-key)
  • If editing: test -f "path/to/input.png"

• Common failures:

  • Error: No API key provided. → set GEMINI_API_KEY or pass --api-key
  • Error loading input image: → wrong path / unreadable file; verify --input-image points to a real image
  • "quota/permission/403" style API errors → wrong key, no access, or quota exceeded; try a different key/account

Filename Generation

Generate filenames with the pattern: yyyy-mm-dd-hh-mm-ss-name.png

Format: {timestamp}-{descriptive-name}.png

• Timestamp: Current date/time in format yyyy-mm-dd-hh-mm-ss (24-hour format)
• Name: Descriptive lowercase text with hyphens
• Keep the descriptive part concise (1-5 words typically)
• Use context from user's prompt or conversation

Examples:

• Prompt "A serene Japanese garden" → 2026-05-16-14-23-05-japanese-garden.png
• Prompt "한복 입은 고양이" → 2026-05-16-15-30-12-hanbok-cat.png

Image Editing

When 지혁 wants to modify an existing image:

1. Check if 지혁 provides an image path or references an image in the current directory
2. Use --input-image parameter with the path to the image
3. The prompt should contain editing instructions
4. Common editing tasks: add/remove elements, change style, adjust colors, blur background, etc.

Prompt Handling

For generation: Pass 지혁's image description as-is to --prompt. 한국어 prompt OK — Gemini는 한국어 처리 가능. Only rework if clearly insufficient.

For editing: Pass editing instructions in --prompt (e.g., "add a rainbow in the sky", "make it look like a watercolor painting")

Preserve 지혁's creative intent in both cases.

Prompt Templates (high hit-rate)

• Generation template:

  • "Create an image of: . Style: . Composition: <camera/shot>. Lighting: <lighting>. Background: <background>. Color palette: <palette>. Avoid: <list>."

• Editing template (preserve everything else):

  • "Change ONLY: . Keep identical: subject, composition/crop, pose, lighting, color palette, background, text, and overall style. Do not add new objects. If text exists, keep it unchanged."

Output

• Saves PNG to current directory (or specified path if filename includes directory)
• Script outputs the full path to the generated image
• Do not read the image back - just inform 지혁 of the saved path
• 텔레그램 전송 필요 시: mcp__send-file__send_photo (10MB 이하; 4K PNG은 sometimes >10MB → JPEG 변환 또는 PDF로 send_document).

When to use this vs existing image-gen

• 기존 image-gen 스킬: OpenAI gpt-image-1 (Codex OAuth 사용). 빠르고 quota 공유.
• 본 스킬 (nano-banana-pro): Google Gemini 3.1 Flash Image. 4K 지원, 다른 스타일, 별도 quota.
• 같은 prompt를 양쪽에 돌려 비교하고 싶을 때 유용.