evolink-nano-banana-2

What This Skill Does

Nano Banana 2 is a streamlined, high-performance image generation skill integrated directly into the OpenClaw agent ecosystem. It serves as an optimized interface for the Google Gemini 3.1 Flash image model, specifically curated for users who prioritize speed, versatility, and ease of use. By leveraging the unified Evolink API, this skill allows you to move seamlessly from text prompts to high-quality visual outputs or perform complex image editing tasks within a single, secure environment. Unlike bulky, all-encompassing media suites, Nano Banana 2 is designed for immediate productivity, stripping away configuration friction while keeping the full power of Gemini 3.1 Flash readily accessible.

Installation

To get started, ensure you have the OpenClaw environment active. Install the skill using the command: clawhub install openclaw/skills/skills/evolinkai/evolink-nano-banana-2. Once installed, you must configure the MCP server to bridge the connection to the Evolink infrastructure. Use the following command for Claude Desktop or Cursor users: npx -y @evolinkai/evolink-media@latest and ensure your EVOLINK_API_KEY is provided in your environment variables. Without this key, the skill will remain in a restricted state, unable to communicate with the generation endpoints. If you encounter setup issues, the skill will automatically provide diagnostics and suggestions to get your server online.

Use Cases

This skill is perfect for rapid prototyping, social media content creation, and conceptual design. Use it to generate high-fidelity assets for presentations, create unique illustrations for blog posts, or edit existing images by uploading them through the File API for in-painting or style modification. Because it utilizes the Gemini 3.1 Flash architecture, it is particularly well-suited for high-volume workflows where waiting minutes for a render is not an option. It functions best as an iterative tool—generate a base image, use the editing tools to refine specific sections, and manage your assets via the file list tool.

Example Prompts

1. "I need a futuristic, minimalist interior design concept for a home office, rendered in a 16:9 aspect ratio with soft, natural lighting."
2. "Please upload my architectural sketch and transform it into a professional, photorealistic render using the Gemini 3.1 Flash model."
3. "Let's try a different mood for the last image; keep the composition the same but shift the color palette to warm sunset tones and make the textures more rustic."

Tips & Limitations

To get the best results, always be specific with your stylistic intent and subject matter. Because this skill follows a 'guide, don't decide' philosophy, treat your initial prompt as a conversation starter. If the model output isn't quite right, use the check_task tool to verify the process and then request variations. Be mindful of the 72-hour expiration window for files uploaded to the server; if you intend to keep your assets long-term, download them to your local storage. Finally, keep your API key secure—it is the master key for your image generation quota and access rights.