LABSCO SUMMARY

Task hand-off between a human's plan and an agent's work, with the trail kept in one place.

The value is the loop rather than any single call: get_next_task decides what to do, start_task and update_progress make that visible on the dashboard while it happens, and create_discussion gives the agent a way out when a requirement is unclear. The plan itself lives on Buildable, so this earns its place only if that is already where your roadmap is.

What it is

An MCP client for Buildable, a hosted build-planning platform. It gives an assistant the project's plan and task list, and a way to write progress back to the same dashboard a human is watching.

What you get

get_project_context — the whole project in one call: plan, tasks and recent activity
get_next_task — the task to work on next, chosen by priority and dependencies
start_task to claim a task, with an optional approach note and a time estimate
update_progress for a percentage and a status line partway through
complete_task to close a task out with notes on what was done
create_discussion to ask a human a question instead of guessing
health_check to confirm the connection before a long run

Requirements

A Buildable account with a project, and an API key generated from that project's settings. The server also needs to know which project it is working on and which assistant is calling; both are supplied through your client config. Runs on Node.

Setup effort

One command plus a key — npx @bldbl/mcp, then supply credentials

See installation options

USE CASES

01

Starting a work session without deciding what to do

Pull the project plan and the next priority task, then claim it before writing any code.

get_project_contextget_next_taskstart_task

02

Keeping a human in the loop on a long task

Report a percentage and a status line as work lands, so the dashboard stays current without anyone asking.

update_progress

03

Asking rather than guessing

Open a discussion on the project when a requirement is ambiguous, then close the task with notes once it is answered.

create_discussioncomplete_task

04

Checking the connection before a long run

Confirm the server can reach Buildable before an agent starts a multi-hour task against it.

health_check

7 tools

Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.

complete_task

create_discussion

get_next_task

get_project_context

health_check

start_task

update_progress

BEFORE YOU INSTALL

What to know first

Tools you get

7 tools when we connected. start_task, update_progress, complete_task and create_discussion change state in your Buildable project; the rest only read.

Requires

npx on PATH
BUILDABLE_API_KEY in your environment
BUILDABLE_PROJECT_ID in your environment

Where it runs

Both documented clients run the server on your machine.

CONTEXT COST

Light Standing cost

Rated by the size of the 7 tool definitions that ride along on every turn, not by how many there are

INSTALLATION

Connect your client

Server setup

TERMINAL

npx @bldbl/mcp

MCP Clients

Claude Desktop — npx

CLAUDE DESKTOP — NPX · JSON

{
  "mcpServers": {
    "buildable": {
      "command": "npx",
      "args": [
        "-y",
        "@bldbl/mcp"
      ],
      "env": {
        "BUILDABLE_API_KEY": "<YOUR_BUILDABLE_API_KEY>",
        "BUILDABLE_PROJECT_ID": "<YOUR_BUILDABLE_PROJECT_ID>",
        "BUILDABLE_AI_ASSISTANT_ID": "claude-desktop"
      }
    }
  }
}

Cursor

CURSOR · JSON

{
  "mcpServers": {
    "buildable": {
      "command": "npx",
      "args": [
        "-y",
        "@bldbl/mcp"
      ],
      "env": {
        "BUILDABLE_API_KEY": "<YOUR_BUILDABLE_API_KEY>",
        "BUILDABLE_PROJECT_ID": "<YOUR_BUILDABLE_PROJECT_ID>",
        "BUILDABLE_AI_ASSISTANT_ID": "cursor-ide"
      }
    }
  }
}

Buildable