← 返回 Skills 市场
748
总下载
2
收藏
1
当前安装
10
版本数
在 OpenClaw 中安装
/install nano-banana-2-skill
功能描述
Generate and edit images using Google's Nano Banana 2 (Imagen) model — the latest high-quality image generation AI. Supports text-to-image generation and ima...
安全使用建议
This skill appears to do what it says: it sends prompts and (with your explicit consent) image files to Atlas Cloud or Google AI Studio and returns generated images. Before installing, consider: (1) the SKILL.md requires you to confirm any local-file uploads — make sure the agent actually asks you before uploading sensitive photos; (2) the registry metadata vs. SKILL.md wording: you only need at least one provider key, not necessarily both — set only the provider key you trust to reduce accidental use; (3) API keys grant billing-capable access — use keys with restricted scope if possible, enable billing alerts, and be prepared to revoke keys; (4) test with non-sensitive images first and review output files before sharing. If you want higher assurance, review the full script locally (it uses only standard-library network calls) or run it in an isolated environment.
功能分析
Type: OpenClaw Skill
Name: nano-banana-2-skill
Version: 1.0.9
The nano-banana-2-skill bundle provides image generation and editing capabilities by interfacing with Atlas Cloud and Google AI Studio APIs. The included Python script (scripts/generate_image.py) is well-implemented using only standard libraries, handles API keys via environment variables, and includes a manual confirmation prompt before uploading local files to the cloud. The SKILL.md instructions correctly guide the agent on provider selection and emphasize user privacy by requiring explicit consent for data uploads, with no evidence of malicious intent, data exfiltration, or prompt injection attacks.
能力评估
Purpose & Capability
The skill claims to call Atlas Cloud or Google AI Studio for image generation and the code + SKILL.md request exactly the corresponding API keys and endpoints. Requiring ATLASCLOUD_API_KEY and/or GEMINI_API_KEY is proportionate to the stated functionality; primaryEnv being ATLASCLOUD_API_KEY is reasonable given Atlas Cloud is the default provider.
Instruction Scope
SKILL.md and the included script limit actions to submitting prompts, uploading local images (with explicit user consent required in the doc), polling for results, and downloading outputs. The instructions do send data to third-party endpoints (api.atlascloud.ai and generativelanguage.googleapis.com) which is expected for this skill and is explicitly called out with a privacy note.
Install Mechanism
There is no install spec (instruction-only plus a small Python script). No downloads or archive extraction are requested. The script uses only Python standard library (urllib), so there is no opaque install-time code execution risk.
Credentials
Requesting Atlas Cloud and Google API keys is justified. Minor inconsistency: registry metadata lists both ATLASCLOUD_API_KEY and GEMINI_API_KEY as 'required env vars', while SKILL.md states 'At least one of the above must be set' and the script supports choosing either provider. This is a documentation/metadata mismatch rather than an unexplained privilege request.
Persistence & Privilege
The skill is not always-included, does not request system-wide configuration changes, and does not persist credentials on its own. Autonomous invocation is allowed by platform default but does not combine with any other broad privileges in this skill.
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-2-skill - 安装完成后，直接呼叫该 Skill 的名称或使用
/nano-banana-2-skill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
v1.0.9
- Added `metadata.openclaw.requires.env` and `primaryEnv` fields to formally declare required environment variables and primary provider.
- No changes to functionality or code; documentation was updated to include structured metadata for environment requirements.
- Ensures better compatibility and clarity for OpenClaw and related tools by explicitly specifying required configuration.
v1.0.8
No changes detected in this release (version 1.0.8); documentation and code remain unchanged.
v1.0.7
nano-banana-2-skill 1.0.7
- Clarified environment variable requirements: ATLASCLOUD_API_KEY is now marked as required unless GEMINI_API_KEY is set.
- Updated env var descriptions for clarity on provider selection and requirement.
- No changes to features, provider logic, or usage instructions.
v1.0.6
**v1.0.6 – Adds Python script for image generation**
- Added `scripts/generate_image.py`, a Python script for listing models, generating images, uploading images, and editing images via the Atlas Cloud API.
- Updated documentation to include usage instructions for the new script.
- No changes to API endpoints or environment variables.
v1.0.5
**v1.0.5 changelog**
- Added explicit privacy & data usage information, including a warning that image data will be uploaded and that user confirmation is required before uploading local files.
- Updated pricing comparison: Atlas Cloud is now shown as 10% cheaper than Google AI Studio for all resolutions.
- Clarified workflow requirements for uploading images: agent must get user approval before any file upload.
- Minor improvements to tables and section formatting for better readability.
v1.0.4
nano-banana-2-skill 1.0.4 changelog:
- Clarified provider selection logic and setup steps for Atlas Cloud and Google AI Studio.
- Simplified and shortened provider comparison and pricing sections.
- Made provider triggers and environment variable requirements more explicit.
- Reduced promotional content and focused documentation on usage, parameters, and API workflows.
- No code or functional changes; documentation cleanup only.
v1.0.3
nano-banana-2-skill v1.0.3 changelog
- Added skill metadata fields: source, homepage, and required environment variables for flexible provider setup.
- Now documents that either ATLASCLOUD_API_KEY or GEMINI_API_KEY is required, depending on provider.
- Clarified provider selection logic: default to Atlas Cloud, but prompt user if both API keys set.
- Documentation reorganized for conciseness and clarity; redundant provider detail and marketing text reduced.
- The skill trigger description no longer recommends one provider over another in the general description.
- No code or functional changes; documentation improvements only.
v1.0.2
**nano-banana-2-skill v1.0.2 changelog**
- Updated SKILL.md with a streamlined, more concise guide focused on practical setup and usage for both Atlas Cloud and Google AI Studio provider modes.
- Added clear Atlas Cloud recommendations (cheaper, unified access, bonus credits) and new user referral link.
- Clarified pricing advantages of Atlas Cloud with detailed, easy-to-read comparison tables.
- Simplified API usage instructions: emphasized upload instructions for editing, cleaner parameter tables, clarified polling workflow, and improved image upload step.
- Removed redundant details and condensed available model listings for better readability.
- Improved guidance and trigger criteria for users requesting image generation or editing.
v1.0.1
update
v1.0.0
release
元数据
常见问题
Nano Banana 2 Image Generation&Editing 是什么？
Generate and edit images using Google's Nano Banana 2 (Imagen) model — the latest high-quality image generation AI. Supports text-to-image generation and ima... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 748 次。
如何安装 Nano Banana 2 Image Generation&Editing？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install nano-banana-2-skill」即可一键安装，无需额外配置。
Nano Banana 2 Image Generation&Editing 是免费的吗？
是的，Nano Banana 2 Image Generation&Editing 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Nano Banana 2 Image Generation&Editing 支持哪些平台？
Nano Banana 2 Image Generation&Editing 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 2 Image Generation&Editing？
由 MikeWang（@xixihhhh）开发并维护，当前版本 v1.0.9。
推荐 Skills