README_en.md•3.54 kB
# Baidu Cloud LLM Security MCP Server
This code repository contains an MCP server that provides security for the content for LLM（ https://cloud.baidu.com/doc/AIGC_SEC/s/qlxblhd5j ）Access to functions.
## Precondition
Before using the Baidu Cloud LLM Security MCP Server, please ensure that you have the following conditions:
1. Python 3.10 or higher version
2. Installed [UV]（ https://github.com/astral-sh/uv ）Used to run MCP Server
## usage
The recommended way to use Baidu Cloud LLM Security MCP Server is to run it through `uv` without the need for installation.
Clone the code repository and execute the following command:
```
git clone https://github.com/liangjunyu2010/mcp_server_safe_content_check.git
cd mcp_server_safe_content_check
```
Subsequently, you can directly run it through `uv`, where `BAIDU_CLOUD_ACCESS_KEY_ID` and `BAIDU_CLOUD_SECRET_ACCESS_KEY` can be modified according to actual needs:
```
uv run src/mcp_server_safe_content_check/server.py
uv run src/mcp_server_safe_content_check/server.py --BAIDU_CLOUD_ACCESS_KEY_ID ACCESS_KEY --BAIDU_CLOUD_SECRET_ACCESS_KEY SECRET_KEY
```
Alternatively, modify the `. env` file in the `src/mcp-user_stafe_ccontent_check/` directory to set environment variables, and then run the server using the following command:
```
uv run src/mcp_server_safe_content_check/server.py
```
## Supported applications
Baidu Cloud LLM Security MCP Server can be used in conjunction with various large language model applications that support model context protocols
- **Cursor**：an artificial intelligence code editor that supports MCP
- **Customize MCP client**：Any application that implements the MCP client specification
## Usage in cursor
[Cursor also supports MCP]（ https://docs.cursor.com/context/model-context-protocol ）Tools. You can add Baidu Cloud LLM Security MCP Server to the cursor in two ways:
依次打开`Cursor Settings`>`MCP`，Click `Add MCP Server`，add config in `mcp.json`：
```JSON
{
"mcpServers": {
"safe-content-check": {
"command": "uv",
"args": [
"run",
"--with",
"mcp[cli]",
"mcp",
"run",
"/PATH/mcp_server_safe_content_check/src/mcp_server_safe_content_check/server.py"
],
"env": {
"BAIDU_CLOUD_ACCESS_KEY_ID": "****",
"BAIDU_CLOUD_SECRET_ACCESS_KEY": "****"
}
}
}
}
```
Restart the Cursor or reload the window.
## Methods
Baidu Cloud LLM Security MCP Server provides the following methods:
### Input detection operation
- `input_analyze`: Detecting input content
- param：
- `text`: The input text content
## Environment variable
[Baidu Cloud IAM](https://console.bce.baidu.com/iam/#/iam/accesslist) Permission `AFDFullControlAccessPolicy`
- `BAIDU_CLOUD_ACCESS_KEY_ID`: Baidu Cloud Authorization ACCESS_KEY
- `BAIDU_CLOUD_SECRET_ACCESS_KEY`: Baidu Cloud Authorization SECRET_KEY
## Using examples
### Cursor
#### Example : Check if there is any unsafe information in the file content
```
帮忙检测下风险
```
The cursor will use the input.analyze provided by the Baidu Cloud LLM Security MCP Server to detect whether the input content is secure
```
检测结果显示，该文本内容存在极高风险：
风险类型：犯罪相关内容（hitType: crime）
风险评分：0.998（满分1分）
处理建议：严禁传播（action: 2）
安全评估：内容不安全（isSafe: 0）
```