Back to Browse

GoalTracker MCP Server

by Nhat-TheDev
Developer ToolsLow Risk8.2LocalNew
Free

Persistent, structured project memory for AI coding agents — Goal → Spec → Milestone → Task.

About

GoalTracker is an MCP (Model Context Protocol) server that gives an AI coding agent (Claude Code, Claude Desktop, or any MCP-compatible client) a real, durable plan to work from instead of an ad-hoc todo list in chat. It stores a `Goal → Spec → Milestone → Task → Note` hierarchy in a single local SQLite file, exposing 14 MCP tools for creating goals, confirming specs, breaking work into milestones and tasks, tracking status with an audit trail, and checkpointing progress.

The core problem it solves is context loss: long-running coding work rarely fits in one session, and a plain chat-based todo list doesn't survive a context reset or a handoff to a different agent. With GoalTracker, a fresh agent session — or an entirely different agent picking up the same work hours or days later — recovers full context (goal, confirmed spec, milestone/task state, and the last checkpoint's summary and next actions) with a single `goal_get_context` call, rather than re-deriving the plan from conversation history.

It is data-only and judgment-free by design: the MCP stores and returns structured data but never makes planning decisions itself. The only two built-in validation gates are (1) a milestone with fewer than 2 active tasks blocks task work until explicitly approved via `milestone_approve`, and (2) `goal_create` requires a non-empty description. Everywhere else, all planning judgment stays with the agent and the user.

It's for developers and teams who use AI coding agents on multi-session or multi-agent projects and want the agent to remember what it's doing, why a task is blocked, and what's left — without re-explaining the plan every time context resets.

Security Report

8.2
Low Risk8.2Low Risk

GoalTracker is a well-structured MCP server for managing project goals and tasks with persistent SQLite storage. Code is clean with proper input validation via Zod schemas, no malicious patterns detected, and no hardcoded credentials. Permissions are appropriate for the stated purpose (file I/O for SQLite database, environment variable for optional path override). Minor findings relate to database error handling and input validation edge cases that do not pose security risks. Package verification found 1 issue.

7 files analyzed · 5 issues found

Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.

Permissions Required

This plugin requests these system permissions. Most are normal for its category.

File System Read

Reads files on your machine. Normal for tools that analyze or process local data.

File System Write

Writes or modifies files on your machine. Check that this is expected for the tool.

env_vars

Check that this permission is expected for this type of plugin.

What You'll Need

Set these up before or after installing:

Path to the SQLite database file. Defaults to `~/.goaltracker/goaltracker.db`, created automatically on first run. Schema migrations apply themselves on startup — no manual migration step is ever needed.Optional

Environment variable: GOALTRACKER_DB_PATH

How to Install

Add this to your MCP configuration file:

{
  "mcpServers": {
    "goaltracker": {
      "env": {
        "GOALTRACKER_DB_PATH": "your-goaltracker-db-path-here"
      },
      "args": [
        "-y",
        "goaltracker"
      ],
      "command": "npx"
    }
  }
}

Getting Started

Once installed, try these example prompts and explore these capabilities:

  • 1"Start tracking a new goal: add dark mode support to the settings page"
  • 2"What's the current status of the goal I was working on?"
  • 3"Resume work — pick up where the last session left off"
  • 4"Give me a status report on this goal before I close it out"
  • 5Tool: `goal_create` — Create a new goal with a required description; the starting point for any tracked piece of work
  • 6Tool: `spec_set` — Record and confirm the spec (overview + acceptance criteria) for a goal before it's broken into milestones
  • 7Tool: `milestone_create` — Break a confirmed spec into milestones
  • 8Tool: `task_create` — Add tasks under a milestone
  • 9Tool: `task_update_status` — Move a task through pending → in_progress → completed/blocked/cancelled, with a required reason for blocked/cancelled
  • 10Tool: `goal_get_context` — One-call recovery of goal, spec, milestones, progress, and last checkpoint; use at the start of any session touching an existing goal
  • 11Tool: `status_report` — Progress snapshot and acceptance-criteria check before closing a goal
  • 12Tool: `checkpoint_save` — Save an agent summary and next actions at any stopping point, for clean handoff to a future session or agent

Reviews

No reviews yet

Be the first to review this server!

GoalTracker MCP Server - Persistent, structured project memory for AI coding agents | MCP Marketplace