WEEEK MCP with a read/write split
weeek-mcp-server
A 12-tool WEEEK MCP server split into 7 read and 5 write tools, so an MCP client can auto-approve the safe ones
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Can create, update and complete tasks in the workspace
- The API token is not read-only by itself; restriction is only possible on the MCP client side
Install
Manual install
{
"mcpServers": {
"weeek": {
"command": "npx",
"args": ["-y", "weeek-mcp-server"],
"env": {
"WEEEK_API_TOKEN": "your-weeek-token-here"
}
}
}
}The README's Claude Desktop config; Cursor uses the same shape in ~/.cursor/mcp.json.
This is third-party code. Review the repository files before installing.
What it does
The server gives access to WEEEK projects, boards, tasks and comments through 12 tools, explicitly split into read (projects, boards, columns, tasks, comments) and write (create, update, move and complete a task, post a comment). The split is deliberate so MCP clients can configure auto-approval for read tools only, leaving writes under manual control. List tools paginate at 20 by default, capped at 50, to stay under the 25k-token MCP response limit. API failures return structured with isError: true and a human-readable message, without crashing the server. It installs and runs via npx with no separate setup.
Who it is for. For people who want their MCP client to auto-approve WEEEK reads while requiring confirmation for writes.
Good fit when
- You want to separate auto-approved reads from confirmation-gated writes in client settings
- You need large task lists without risking the MCP response limit
- You need to work with task comments alongside the tasks themselves
Not a fit when
- You need WEEEK's knowledge base or CRM parts, the server only covers tasks
- You need a CLI alongside the MCP, this ships only the server
Example request
Show my open tasks in the Website project and move task 42 to the Done columnLimitations
The server does not surface its description in the repository card (no repoData description); it needs a WEEEK API token with full workspace rights. Coverage is limited to tasks, boards and comments.
How to disable. Remove the weeek server from your MCP client configuration and revoke the token in WEEEK settings.
MCP
- Transport
- stdio
- Authentication
- API key
| Environment variables | |
|---|---|
| WEEEK_API_TOKEN required, secret | A personal WEEEK token, granting full read and write access to the workspace. |
Security check
- Can create, update and complete tasks in the workspace
- The API token is not read-only by itself; restriction is only possible on the MCP client side
README in short
The README lists 12 tools with an explicit read/write split and explains the motivation: MCP clients can auto-approve reads and require confirmation for writes. It describes safe pagination defaults, a single WEEEK_API_TOKEN variable, structured error handling, and configs for Claude Desktop and Cursor via npx with no install.
FAQ
Can I restrict the agent to WEEEK reads only?
Yes, tools are explicitly split into read and write, so the client can be configured to auto-approve only the read set.
What happens if a task list is too large?
List tools paginate at 20 items by default, capped at 50, so the response stays under the MCP 25k-token limit.
Related
A self-hosted knowledge base with block-level references and a built-in MCP server for connecting AI agents to your notes
A CLI for every Google Workspace API with JSON output and agent skills: Drive, Gmail, Calendar, Sheets and more
Local search over Markdown notes, docs and meeting transcripts: keywords, semantic search and reranking, with an MCP server
A task manager for AI-driven development: breaks a PRD into dependent tasks and guides the agent through them via MCP or CLI