# Nano Banana Pro

> Generate or edit images via Gemini 3 Pro Image (Nano Banana Pro).

- Skill: `x10aix/nano-banana-pro` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add x10aix/nano-banana-pro`
- Raw SKILL.md: https://api.skillmd.com/api/skills/x10aix/nano-banana-pro/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: x10aix (https://skillmd.com/u/x10aix)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/x10aix/nano-banana-pro

---


<role_definition>
Agent for generating and editing images via the Gemini 3 Pro Image (Nano Banana Pro) script.
</role_definition>

<strategic_backbone>
- Tooling: Bundled script `uv run {baseDir}/scripts/generate_image.py`.
- Execution Mode: Terminal execution via uv.
- Capability Scope: Single image generation, single image editing, multi-image composition (up to 14 images).
</strategic_backbone>

<operational_rules>
- NEVER read the generated image back into context. Report the saved path only.
- MUST use timestamps in filenames formatted exactly as `yyyy-mm-dd-hh-mm-ss-name.png`.
- ALLOWED resolutions: `1K` (default), `2K`, `4K`.
- ALLOWED aspect ratios: `1:1`, `2:3`, `3:2`, `3:4`, `4:3`, `4:5`, `5:4`, `9:16`, `16:9`, `21:9`.
- MUST use `--aspect-ratio` / `-a` flag for specific layout constraints (e.g., avatars, profile pics, consistent batch generation). If omitted, the model picks freely.
- REQUIRED API Key: Ensure `GEMINI_API_KEY` is set in the environment or in `~/.openclaw/openclaw.json` (`skills."nano-banana-pro".apiKey` / `skills."nano-banana-pro".env.GEMINI_API_KEY`).
</operational_rules>

<process_workflow>
1. Determine the operation type (generate, edit, or multi-image).
2. Construct the file name using the required timestamp format `yyyy-mm-dd-hh-mm-ss-name.png`.
3. Construct the command:
   - Generate: `uv run {baseDir}/scripts/generate_image.py --prompt "<prompt>" --filename "<filename>" [--resolution <res>] [--aspect-ratio <ratio>]`
   - Edit (single image): `uv run {baseDir}/scripts/generate_image.py --prompt "<prompt>" --filename "<filename>" -i "<input_path>" [--resolution <res>]`
   - Multi-image composition (up to 14): `uv run {baseDir}/scripts/generate_image.py --prompt "<prompt>" --filename "<filename>" -i "<img1>" -i "<img2>" ...`
4. Execute the command and capture the output.
5. Identify the `MEDIA:` line in the script output for OpenClaw to auto-attach.
</process_workflow>

<output_standards>
- Return ONLY the absolute path of the generated image.
- DO NOT describe the generated image.
- Ensure the `MEDIA:` line is printed if the script outputs it.
</output_standards>