home / skills / openclaw / skills / minimax-understand-image

minimax-understand-image skill

/skills/thincher/minimax-understand-image

This skill analyzes images using MiniMax MCP to identify objects, text, and scenes, delivering descriptive insights and actionable image understanding.

npx playbooks add skill openclaw/skills --skill minimax-understand-image

Review the files below or copy the command above to add this skill to your agents.

Files (3)
SKILL.md
3.9 KB
---
name: minimax-understand-image
description: 使用 MiniMax MCP 进行图像理解和分析。触发条件:(1) 用户要求分析图片、理解图像、描述图片内容 (2) 需要识别图片中的物体、文字、场景 (3) 使用 MiniMax 的 understand_image 功能
---

# minimax-understand-image

使用 MiniMax MCP 服务器进行图像理解和分析。

## 执行流程(首次需要安装,后续直接步骤4调用)

### 步骤 1: 检查并安装依赖

#### 1.1 检查 uvx 是否可用

```bash
which uvx
```

如果不存在,安装 uv:

**方法 1: 使用官方安装脚本(推荐)**
```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```

**方法 2: 使用国内镜像加速(如果官方脚本下载失败)**

临时使用清华镜像源安装:
```bash
export UV_INDEX_URL="https://pypi.tuna.tsinghua.edu.cn/simple"
curl -LsSf https://astral.sh/uv/install.sh | sh
```

或者临时使用阿里云镜像源:
```bash
export UV_INDEX_URL="https://mirrors.aliyun.com/pypi/simple/"
curl -LsSf https://astral.sh/uv/install.sh | sh
```

#### 1.2 检查 MCP 服务器是否已安装

```bash
uvx minimax-coding-plan-mcp --help
```

执行命令判断是否MCP服务器已安装, 如果安装了跳到步骤 2。

#### 1.3 安装 MCP 服务器(如果未安装)

**方法 1: 使用默认源安装**
```bash
uvx install minimax-coding-plan-mcp
```

**方法 2: 使用国内镜像加速(如果默认源下载失败)**

临时使用清华镜像源:
```bash
export UV_INDEX_URL="https://pypi.tuna.tsinghua.edu.cn/simple"
uvx install minimax-coding-plan-mcp
```

或者临时使用阿里云镜像源:
```bash
export UV_INDEX_URL="https://mirrors.aliyun.com/pypi/simple/"
uvx install minimax-coding-plan-mcp
```

### 步骤 2: 检查 API Key 配置

```bash
cat ~/.openclaw/config/minimax.json 2>/dev/null | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('api_key', ''))"
```

如果返回非空的 API Key,跳到步骤 4。

### 步骤 3: 配置 API Key(如果未配置)

#### 3.1 尝试从 ~/.openclaw/agents/main/agent/auth-profiles.json 中的配置文件中获取

根据返回的判断:
- 名称包含 "minimax" 或 "MiniMax"

找到匹配的 Key 后,询问用户确认是否使用。

#### 3.2 如果没有找到 Key,向用户索要

直接询问用户提供 MiniMax API Key。
如果未购买MiniMax,购买地址为: https://platform.minimaxi.com/subscribe/coding-plan?code=GjuAjhGKqQ&source=link

#### 3.3 保存 API Key

```bash
mkdir -p ~/.openclaw/config
cat > ~/.openclaw/config/minimax.json << EOF
{
  "api_key": "API密钥",
  "output_path": "~/.openclaw/workspace/minimax-output"
}
EOF
```



### 步骤 4: 使用 MCP 处理图像

#### 4.1 准备图片

将图片放到可访问路径,例如:
- `~/.openclaw/workspace/images/图片名.jpg`
- 或者使用 URL

#### 4.2 调用 understand_image

使用脚本调用 MCP 服务:

```bash
python3 {curDir}/scripts/understand_image.py <图片路径或URL> "<对图片的提问>"
```

**示例:**

```bash
# 描述图片内容
python3 {curDir}/scripts/understand_image.py ~/image.jpg "详细描述这张图片的内容"

# 使用 URL
python3 {curDir}/scripts/understand_image.py "https://example.com/image.jpg "这张图片展示了什么?"
```

#### 4.3 API 参数说明

| 参数 | 说明 | 类型 |
|------|------|------|
| image | 图片路径或 URL | string (必填) |
| prompt | 对图片的提问 | string (必填) |


## 脚本说明

脚本位置:`{curDir}/scripts/understand_image.py`

**功能:**
- 优先从环境变量 `MINIMAX_API_KEY` 读取 API Key,如果没有则从 `~/.openclaw/config/minimax.json` 读取
- 通过 stdio 模式启动 MCP 服务器
- 发送 JSON-RPC 请求调用 `understand_image` 工具
- 返回格式化的 JSON 结果

**错误处理:**
- API Key 未配置时提示错误
- uvx 未安装时提示安装命令
- MCP 服务器错误时显示 stderr 输出

Overview

This skill uses the MiniMax MCP server to perform image understanding and analysis. It provides a lightweight command-line workflow to call the understand_image tool and receive structured JSON descriptions. The skill automates dependency checks, API key handling, and the call to the MCP service for fast image inspection.

How this skill works

The skill checks for the uvx runtime and installs the minimax-coding-plan-mcp package if missing. It validates or prompts for a MiniMax API key stored in ~/.openclaw/config/minimax.json. Finally, it invokes a Python helper script that sends a JSON-RPC request to the MCP server's understand_image tool and returns formatted JSON results.

When to use it

  • You need a detailed description of a photo, scene, or object in an image.
  • You must extract visible text, identify objects, or summarize visual context.
  • You want to automate image analysis via CLI or scriptable pipelines.
  • You prefer using a local MCP runtime with a configured MiniMax API key.

Best practices

  • Ensure uvx is installed before running the skill; follow the provided install commands if missing.
  • Store a valid MiniMax API key in ~/.openclaw/config/minimax.json using the exact JSON structure.
  • Provide clear, focused prompts when calling understand_image to get targeted answers.
  • Use accessible file paths or public URLs for images so the MCP server can fetch them.
  • Check MCP server stderr output for troubleshooting if the call fails.

Example use cases

  • Describe the contents of a product photo to generate alt text for accessibility.
  • Detect and transcribe visible text in a screenshot or document image.
  • Identify objects and scene context in surveillance or field photos.
  • Integrate into a preprocessing pipeline that tags images with semantic labels.
  • Quickly validate that an image contains expected elements before downstream processing.

FAQ

What image formats and sources are supported?

You can use local image paths (e.g., /root/path/image.jpg) or public URLs. Common formats like JPEG and PNG are supported; ensure the MCP server can access the file or URL.

How do I provide or update the MiniMax API key?

Save the key in ~/.openclaw/config/minimax.json with the fields api_key and output_path. The script also attempts to discover keys from gateway providers and will ask you to confirm before saving.