Hong Kong Government Development, Geography and Land Information MCP Server
# Hong Kong Government Development, Geography and Land Information MCP Server
[](https://github.com/hkopenai/hk-development-mcp-server)
[](https://opensource.org/licenses/MIT)
This is an MCP server that provides access to government development, geography and land information data through a FastMCP interface.
## Features
### New Buildings - Plans Processed
- Retrieve data on the number of plans processed by the Building Authority in Hong Kong for new buildings within a specified year range.
## Data Source
- Building plan data from Buildings Department
## Examples
* Get data on new building plans processed by the Building Authority
## Setup
1. Clone this repository
2. Install Python dependencies:
```bash
pip install -r requirements.txt
```
3. Run the server:
```bash
python server.py
```
### Running Options
- Default stdio mode: `python server.py`
- SSE mode (port 8000): `python server.py --sse`
## Cline Integration
To connect this MCP server to Cline using stdio:
1. Add this configuration to your Cline MCP settings (cline_mcp_settings.json):
```json
{
"hk-development": {
"disabled": false,
"timeout": 3,
"type": "stdio",
"command": "uvx",
"args": [
"hkopenai.hk-development-mcp-server"
]
}
}
```
## Testing
Tests are available in the `tests/` directory. Run with:
```bash
pytest
```
TDQS
Scored across 1 tool
With only one tool in the set, there is no possibility of an agent confusing it with another tool. Its purpose—retrieving new building plan processing counts—is clearly distinct by default.
The single tool follows a clear verb_noun snake_case pattern (get_new_building_plans_processed). With one tool there is no inconsistency to introduce, and the name is descriptive and readable.
A server framed as covering Hong Kong development, geography, and land information exposes only one narrow statistic. This is an extreme under-scoping: an agent needing land, geographic, or planning data has essentially no surface to work with.
The domain implied by the server name (development, geography, land information) suggests many data families—land parcels, zoning, geography, plan approvals—yet only one time-series metric is available. There is no lookup, search, or complementary operation, leaving severe gaps.