mcp-vibekanban
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_PORT | No | Port for HTTP transport. | 3000 |
| VIBE_API_URL | No | Vibe Kanban API base URL. | http://localhost:9119 |
| VIBE_REPO_ID | No | Repo UUID. Auto-fetched if project has exactly one repo. | auto-detected |
| MCP_TRANSPORT | No | Transport type: stdio or http. | stdio |
| VIBE_PROJECT_ID | Yes | Required. UUID of the project to lock onto. | |
| VIBE_WORKSPACE_ID | No | Enables vibe://context resource and get_context tool. | |
| VIBE_RESOURCE_POLL_INTERVAL | No | Resource subscription poll interval (ms). | 10000 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tasks | {
"list": {},
"cancel": {},
"requests": {
"tools": {
"call": {}
}
}
} |
| tools | {} |
| resources | {
"subscribe": true,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_contextA | Get current workspace context. Check active project, task, and workspace info. get_context() |
| list_tasksA | List tasks in the project. View all tasks or filter by status (todo, inprogress, inreview, done, cancelled). list_tasks(status="inprogress", limit=10) |
| create_taskA | Create a new task. Add a task to the project. Supports @tagname expansion in description. create_task(title="Add user auth", description="Implement OAuth with @google-auth") |
| get_taskA | Get task details. View full task information including description. get_task(task_id="abc123...") |
| update_taskB | Update task properties. Change task title, description, or status. update_task(task_id="abc123...", status="done") |
| delete_taskA | Delete a task permanently. Remove a task from the project. Cannot be undone. delete_task(task_id="abc123...") |
| start_workspace_sessionB | Start a coding session for a task. Launch a coding agent to work on a task. Creates workspace + session. start_workspace_session(task_id="abc123...", executor="claude_code") |
| list_sessionsA | List sessions in a workspace. See all executor sessions for a workspace. Each session tracks conversation with one executor. list_sessions(workspace_id="xyz789...") |
| get_sessionA | Get session details. Check which executor is assigned and session state. get_session(session_id="sess123...") |
| send_messageA | Send a message to a coding agent session. Send follow-up instructions to an active session. Auto-queues if executor is busy. send_message(session_id="sess123...", prompt="Add error handling") |
| get_queue_statusB | Check if a message is queued for a session. See pending message when executor is busy. get_queue_status(session_id="sess123...") |
| cancel_queueB | Cancel a queued message. Remove pending message from queue. cancel_queue(session_id="sess123...") |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Project Tasks | All tasks in the current project |
TDQS
Scored across 12 tools
Each tool targets a distinct action and resource. Task tools (create, read, update, delete) are clearly separated from session tools (start, list, get, send, cancel, queue status) and context retrieval. No overlap or ambiguity.
All tools follow a consistent verb_noun pattern in snake_case (e.g., create_task, list_sessions, get_queue_status). The verbs are descriptive and the pattern is uniform across the entire set.
With 12 tools, the server covers task management (CRUD), session lifecycle, queue operations, and workspace context. The count feels well-balanced for the domain without being excessive or sparse.
Task CRUD is fully covered. Sessions have start, list, get, and message sending. Lacks an explicit end/close session tool, but the current set supports core workflows. Minor gap, but agents can manage via cancel_queue or context.