← 返回 Skills 市场

Ai Image Gen Skill

Name: Ai Image Gen Skill
Author: dongrebeccahhh-boop

作者 dongrebeccahhh-boop · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

168

总下载

当前安装

版本数

在 OpenClaw 中安装

/install ai-image-gen-skill

功能描述

AI图片生成技能，使用Gemini生成高质量图片，支持文生图和图生图

使用说明 (SKILL.md)

AI图片生成技能

使用 AI 生成高质量图片。

核心命令

uv run /usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py \
  --prompt "你的描述" --filename "output.jpg" --resolution 2K

参数说明

参数	说明
`--prompt`	图片描述（中英文都行）
`--filename`	输出文件名
`--resolution`	分辨率：1K / 2K
`-i reference.jpg`	可选：参考图片（风格迁移、图片编辑）

使用示例

文生图

# 生成产品封面
uv run /usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py \
  --prompt "极简白色背景产品展示图，现代科技感" \
  --filename cover.jpg \
  --resolution 2K

图生图（风格迁移）

# 基于参考图修改
uv run /usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py \
  --prompt "把背景改成蓝天白云" \
  -i original.jpg \
  --filename edited.jpg

常用场景

# 公众号封面
--prompt "公众号封面：AI技术，科技感，简洁大气"

# 产品图
--prompt "产品展示图，白色背景，专业摄影风格"

# 插图
--prompt "文章配图：数据分析，图表，商务风格"

注意事项

需要 GEMINI_API_KEY 环境变量
支持 1K 和 2K 分辨率
中英文 prompt 都支持

技能创建时间: 2026-03-17

安全使用建议

Do not install or run this skill as-is. Key issues: SKILL.md instructs running an external script at /usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py that is not included, and it requires GEMINI_API_KEY even though metadata lists no env vars. Before proceeding, ask the author for: (1) the actual code or an install spec that creates the referenced script path, (2) an explanation why the skill points to 'nano-banana-pro' instead of its own files, and (3) updated metadata declaring GEMINI_API_KEY. If you must test, inspect the referenced generate_image.py file contents and the 'uv' runner first, run in a sandboxed environment, and ensure the Gemini API key you provide follows least-privilege practices (rotate it, limit scope). Prefer skills with a verifiable source/homepage and included install steps or code.

功能分析

Type: OpenClaw Skill Name: ai-image-gen-skill Version: 1.0.0 The skill instructions in SKILL.md reference a hardcoded execution path (/usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py) that is inconsistent with the skill's own metadata slug (ai-image-gen-skill). Additionally, the actual Python implementation script is missing from the bundle, and the use of a hardcoded path to a different skill's directory is a highly irregular practice that could be used to trigger unintended code execution or cross-skill interference.

能力评估

⚠ Purpose & Capability

The skill claims 'AI图片生成' with Gemini, which fits needing a GEMINI_API_KEY, but the SKILL.md points to /usr/lib/node_modules/openclaw/skills/nano-banana-pro/scripts/generate_image.py — a script that is not part of this skill package and is from a differently named skill (nano-banana-pro). The external dependency is not declared in metadata.

⚠ Instruction Scope

Instructions tell the agent to run an absolute path to a local Python script (via 'uv run') and to read reference image files; they also state a required GEMINI_API_KEY. The metadata declares no env vars and no code files. The instructions therefore ask the agent to execute and access resources outside the skill's own bundle and to use an undeclared secret.

⚠ Install Mechanism

There is no install spec (instruction-only), which is normally low risk — but here the instructions rely on a specific script path under /usr/lib/node_modules/... that this skill did not install. That could cause execution of arbitrary local code if the referenced script exists and is malicious or unexpected.

⚠ Credentials

SKILL.md says '需要 GEMINI_API_KEY' but the skill metadata lists no required env vars or primary credential. Requesting an API key for Gemini is reasonable for an image generation skill, but the omission from declared requirements is an incoherence that prevents informed permissioning and review.

✓ Persistence & Privilege

The skill is not always-enabled and uses normal autonomous invocation defaults. It does not request persistent system-wide privileges in the metadata.

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 ai-image-gen-skill
安装完成后，直接呼叫该 Skill 的名称或使用 /ai-image-gen-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

v1.0.0

ai-image-gen-skill 1.0.0 初始发布 - 新增AI图片生成能力，支持文本生成图片（文生图）和图片参考生成图片（图生图/风格迁移） - 支持中文和英文通过prompt描述图片内容 - 提供分辨率选项：1K和2K - 需配置 GEMINI_API_KEY 环境变量 - 提供详细命令行参数说明和多场景使用示例

元数据

Slug ai-image-gen-skill

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题