MCP01: Google HTML Fetcher MCP Server
# MCP01: Google HTML Fetcher MCP Server
A Model Context Protocol (MCP) server that connects to `google.com`, retrieves the page's HTML source, and stores it in your local directory.
---
## Features
- **Standard MCP Protocol**: Communicates using JSON-RPC 2.0 over standard I/O (`stdio`).
- **Webpage Fetching**: Connects to `google.com` (or user-defined target URL) with HTTP status inspection.
- **Local Persistence**: Writes the HTML response directly to a designated local path (defaults to `C:\Users\hiren\Documents\mcpserver\mcp01\google.html`).
- **Cloud-Ready / NPM Executable**: Can be installed or run on-demand directly from GitHub via `npx` or `npm`.
---
## Exposed MCP Tools
### `fetch_google_html`
Fetches HTML source from `google.com` and writes the resulting content to disk.
#### Input Schema
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
| `url` | string | No | `https://www.google.com` | Target URL to fetch |
| `filename` | string | No | `google.html` | Destination filename |
---
## Installation & Usage
### Option 1: Run & Add Directly in Antigravity CLI (`agy`) via GitHub
You can register this MCP server in `agy` directly from GitHub using `npx`:
```bash
agy mcp add mcp01 npx -y github:hirenputu/mcp01-google-fetcher
```
Check status anytime:
```bash
agy mcp list
```
### Option 2: Run via NPM from GitHub
Run on-demand:
```bash
npx -y github:hirenputu/mcp01-google-fetcher
```
Or install globally:
```bash
npm install -g github:hirenputu/mcp01-google-fetcher
mcp01
```
### Option 3: Local Development
Clone this repository and run locally:
```bash
npm install
node index.js
```
---
## Environment Configuration
| Variable | Description | Default |
|---|---|---|
| `MCP_OUTPUT_DIR` | Directory where HTML files will be saved | `C:\Users\hiren\Documents\mcpserver\mcp01` |
---
## Permissions & Security
- **Strict Access Control**: No personal credentials, tokens, or private secrets are tracked in this repository.
- **Security Policy**: For reporting vulnerabilities, consult [SECURITY.md](SECURITY.md).
- **Network Boundaries**: Only outbound HTTP GET requests are performed.
---
## License
This project is licensed under the [PolyForm Noncommercial License 1.0.0](LICENSE) - see the [LICENSE](LICENSE) file for details.
TDQS
Scored across 1 tool
With only one tool in the set, there is no possibility of overlap or misselection. The single tool's purpose (fetch google.com HTML and save it to disk) is unambiguous.
The lone name fetch_google_html follows a clean snake_case verb_noun pattern. There are no other names to conflict with, so consistency is trivially satisfied.
A single tool is borderline thin even for a narrowly scoped fetcher, and it bundles fetching plus file storage into one opaque operation. The scope is small enough that one tool is defensible, but it feels under-built.
The tool hardcodes google.com and a local Windows path, and there is no companion tool to read, list, or delete the stored HTML, so an agent cannot retrieve the fetched content. Coverage of even this tiny domain has a dead end at the output stage.