← 返回 Skills 市场
Nano Banana Pro
作者
Peter Steinberger
· GitHub ↗
· v1.0.1
103871
总下载
406
收藏
1897
当前安装
2
版本数
在 OpenClaw 中安装
/install nano-banana-pro
功能描述
Generate/edit images with Nano Banana Pro (Gemini 3 Pro Image). Use for image create/modify requests incl. edits. Supports text-to-image + image-to-image; 1K/2K/4K; use --input-image.
安全使用建议
Install only if you are comfortable sending image prompts and any selected input images to Google's Gemini service. Do not use screenshots, documents, credentials, private photos, regulated data, or other sensitive content as input images unless you accept that external processing.
功能分析
Type: OpenClaw Skill
Name: nano-banana-pro
Version: 1.0.1
The `scripts/generate_image.py` script allows reading arbitrary local files specified via the `--input-image` argument using `PIL.Image.open()`. If the provided path points to a sensitive file (e.g., a screenshot of credentials, or a document containing PII) that PIL can successfully open, its content will be sent to the external Google Gemini API (`gemini-3-pro-image-preview`). While the stated purpose is image editing, this capability presents a significant data exfiltration risk if the agent is prompted to provide paths to sensitive files, even without explicit malicious intent from the skill developer.
能力评估
Purpose & Capability
The script and instructions are coherent with the stated image generation/editing purpose, including optional user-selected input images and Gemini API use.
Instruction Scope
The instructions are user-directed and scoped to running the image script, but the privacy implications of sending prompts and input images to Google could be clearer.
Install Mechanism
No unusual installer, hidden package behavior, or unrelated setup mechanism was found; the script uses inline uv dependencies for google-genai and pillow.
Credentials
Use of GEMINI_API_KEY, network API calls, and reading an explicitly supplied image path are proportionate for this skill, though sensitive files should not be used as input images.
Persistence & Privilege
No persistence, background execution, privilege escalation, credential harvesting, destructive behavior, or broad local indexing was found.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install nano-banana-pro - 安装完成后，直接呼叫该 Skill 的名称或使用
/nano-banana-pro触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.1
v1.0.0
元数据
常见问题
Nano Banana Pro 是什么？
Generate/edit images with Nano Banana Pro (Gemini 3 Pro Image). Use for image create/modify requests incl. edits. Supports text-to-image + image-to-image; 1K/2K/4K; use --input-image. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 103871 次。
如何安装 Nano Banana Pro？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install nano-banana-pro」即可一键安装，无需额外配置。
Nano Banana Pro 是免费的吗？
是的，Nano Banana Pro 完全免费（开源免费），可自由下载、安装和使用。
Nano Banana Pro 支持哪些平台？
Nano Banana Pro 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Nano Banana Pro？
由 Peter Steinberger（@steipete）开发并维护，当前版本 v1.0.1。
推荐 Skills