image-generatorlisted

## AI Image Generation Calls Gemini API via the `scripts/generate-image.sh` script to generate images. ### Usage ```bash bash {SKILL_DIR}/scripts/generate-image.sh "<English prompt>" ``` The script will: 1. Automatically obtain Gemini API Key from openteam-server (via `OPENTEAM_TOKEN` env var) 2. Call Gemini generateContent API 3. Save the generated image to `~/.openteam/images/` directory 4. Output image path marker `[generated_image:<path>]` After generation, use the Read tool to view the image and verify the result. ### Prompt Optimization Principles User descriptions are often brief — supplement necessary details like a professional photographer/artist: - Subject description (subject & composition) - Style specification (photorealistic / anime / watercolor / oil painting / digital art, etc.) - Lighting and atmosphere (lighting, mood, atmosphere) - Quality requirements (high detail, 4K, professional) - Angle and composition (close-up, wide shot, bird's eye view, etc.) ### Example User says "draw a cat": ```bash bash {SKILL_DIR}/scripts/generate-image.sh "A fluffy orange tabby cat lounging on a sunlit windowsill, soft bokeh background with indoor plants, warm golden hour lighting, photorealistic, high detail" ``` ### Notes - Each call generates one image only - Generation takes approximately 10-30 seconds - English prompts produce the best results - No NSFW, violent, or inappropriate content - On failure, analyze the reason and suggest prompt modifications