ai-image
AI image generation and editing.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
AI image generation and editing.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
| name | ai-image |
| description | AI image generation and editing. |
| metadata | {"tags":["aigc","gcp"]} |
tldr: Run uv run ${SKILL_DIR}/scripts/generate_image.py Image Generation.
When generating images for a specific persona or character:
--image flag.--prompt, focus on describing the new environment, lighting, pose, and clothing. The model will use the reference image to maintain the facial features and identity.Below are the most common use cases. For full parameters information, run uv run ${SKILL_DIR}/scripts/generate_image.py --help.
Text-to-Image:
--prompt "A hyperrealistic render of a neon jellyfish floating in a cyber forest" --output "neon_jellyfish.png" --aspect-ratio "9:16"
Image-to-Image (Character consistency):
--image "path/to/character_face.jpg" --prompt "The same woman sitting on a cozy sofa at night, wearing a dark oversized hoodie, looking slightly annoyed, cinematic lighting" --output "character_on_sofa.png" --aspect-ratio "3:4"
For errors encountered during API calls, please check the API reference to find the correct parameters.