Cosense MCP Server
# cosense-mcp-server MCP Server
MCP server for [cosense](https://cosen.se)
## Features
- Get Page
## Development
Install dependencies:
```bash
npm install
```
Build the server:
```bash
npm run build
```
For development with auto-rebuild:
```bash
npm run watch
```
## Installation
```bash
git clone https://github.com/funwarioisii/cosense-mcp-server.git
cd cosense-mcp-server
npm run install
npm run build
```
To use with Claude Desktop, add the server config:
On MacOS: `~/Library/Application\ Support/Claude/claude_desktop_config.json`
On Windows: `%APPDATA%/Claude/claude_desktop_config.json`
```json
{
"mcpServers": {
"cosense-mcp-server": {
"command": "node",
"args": ["/path/to/cosense-mcp-server/build/index.js"],
"env": {
"COSENSE_PROJECT_NAME": "your_project_name",
"COSENSE_SID": "your_sid"
}
}
}
}
```
`COSENSE_SID` is optional.
If you want to use this server towards a private project, you need to set `COSENSE_SID`.
### Debugging
Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the [MCP Inspector](https://github.com/modelcontextprotocol/inspector), which is available as a package script:
```bash
npm run inspector
```
The Inspector will provide a URL to access debugging tools in your browser.
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: 'get_page' retrieves a specific page with its content and links, while 'list_pages' enumerates all known pages. There is no overlap or ambiguity between them, as one is for detailed retrieval and the other for listing.
Both tools follow a consistent verb_noun pattern with snake_case: 'get_page' and 'list_pages'. The naming is predictable and readable, with no deviations or mixed conventions, making it easy to understand their functions at a glance.
With only 2 tools, the server feels too thin for its apparent domain of managing pages in a project. It lacks essential operations like creating, updating, or deleting pages, which are typical for such a domain, making it under-scoped and limiting for agents.
The tool surface is severely incomplete for page management. While it provides read operations (get and list), it misses critical CRUD functions such as create, update, and delete, as well as other potential operations like searching or managing links. This will likely cause agent failures in handling full page lifecycles.