Google Banana（Imagen系列）AI文生图提示词生成技能。用于生成高质量角色图、场景图、分镜素材图、概念图的英文Prompt，同时提供中文说明、负面提示词和参数建议。触发词：Banana提示词、Banana作图、Google图片提示词、分镜素材图提示词。

This skill is internally coherent and low-risk from a security/privilege perspective: it only contains prompt templates and examples and asks for no credentials or installs. Before using, consider these practical points: (1) Do not paste private or sensitive data into prompts you generate — the skill only produces text and may be sent to a model you use. (2) Check content policies of the image model you ultimately run (avoid generating prompts that produce illegal or disallowed content). (3) If you want stronger provenance, prefer skills with a recognized source/homepage or author; this skill has unknown source metadata. (4) Because it can be invoked by the agent, review any agent automation rules so it isn't invoked in surprising contexts, but that is a platform default and not specific to this skill.

Type: OpenClaw Skill Name: eleven-banana-prompt-generator Version: 1.0.0 The skill bundle is a legitimate prompt engineering tool designed to help users generate high-quality image prompts for Google's Imagen (Banana) model. The files (SKILL.md and references/banana-prompt-template.md) contain only instructional text, templates, and keyword libraries for various visual styles and compositions, with no executable code, network requests, or suspicious prompt injection attempts.