Is this MCP actually malicious?
No. The malicious behavior is simulated and has been restricted to harmless testing only. The data is sent to a public test service and will not be stored or misused.
@cluster3115
基本信息
If you are a Platform Reviewer:
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"weather-mcp-server": {
"command": "uv run /path/to/weather-mcp-server",
"env": {
"WEATHER_API_KEY": "your-api-key"
}
}
}
}工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。
Simpleweatherforecastserver is a weather forecast server built with the FastMCP framework that provides global weather forecasts and current conditions. However, it is a security testing tool designed to simulate malicious data collection behavior—specifically, an attempt to exfiltrate the user's OpenWeatherMap API key and location. Its actual code is harmless and intended to test a platform’s ability to review and block potentially dangerous MCPs.
Install with Python and dependencies. Run locally using uv run /path/to/weather-mcp-server with the environment variable WEATHER_API_KEY set to a key from WeatherAPI.com. Alternatively, run in remote mode by specifying a URL like http://host:port/sse in your MCP client configuration.
No. The malicious behavior is simulated and has been restricted to harmless testing only. The data is sent to a public test service and will not be stored or misused.
It simulates collecting the user’s OpenWeatherMap API key and geographic location information from the weather query.
It is intended for platform reviewers and security researchers to test whether a review mechanism can detect and block suspicious MCPs. Users should not run it on an unapproved platform.
You need Python, the `uv` tool, and the FastMCP framework. An API key from WeatherAPI.com is required for the local configuration.
Yes, for fetching live weather data and for the simulated data exfiltration to a public HTTP test endpoint.
Open-source MCP to use Blender with any LLM

See and control a real Windows PC you own — from any MCP client, locally or remotely. Observe (UIA + screenshots), click/type/drag/scroll, launch apps, owner Live View. BYOH: your machine, your key.
Hindsight: Agent Memory That Learns

B2B competitive intelligence, live sales battlecards, 5-axis ICP prospect fit scoring, and tech stack teardowns over MCP and REST with cited public-web evidence.
PickFu MCP is a Model Context Protocol (MCP) server that connects AI assistants, developers, marketers, product managers, and AI agents to PickFu's consumer research platform. It enables users to create surveys, conduct
评论