Vincent Carpenter has released GSD Task Manager v13.9.2, a privacy-first productivity tool that now includes a standalone MCP server. This addition allows AI agents to interact directly with a user’s task list, enabling structured decision-making through the Eisenhower matrix. The project, hosted on GitHub, emphasizes local data custody while offering optional cloud sync, marking a significant step in integrating autonomous agents with personal workflow systems.

The MCP Integration

The core of this update is the Model Context Protocol server, which exposes 19 distinct tools for task management, analytics, and diagnostics. This allows AI agents to not only read tasks but also perform validated and dry-run-aware writes. By providing a structured interface for an Eisenhower matrix, the system enables agents to categorize work into Do First, Schedule, Delegate, or Eliminate quadrants, effectively giving them a logical framework for prioritization rather than just a flat list.

Privacy and Architecture

GSD maintains a strict privacy-first posture, defaulting to local-only storage in the browser’s IndexedDB. Cloud synchronization is entirely optional, utilizing PocketBase for multi-device support with OAuth integration for Google, Apple, or GitHub. The application is built on Next.js 16 and React 19, ensuring a client-side static export that can be deployed without a backend. This architecture ensures that users retain full control over their data, with clear documentation on security trust boundaries and token storage.

Key Takeaways

  • The MCP server enables AI agents to manage tasks using validated writes and dry-run modes.
  • Data privacy is prioritized with local IndexedDB storage and optional, self-hostable PocketBase sync.
  • The tool implements the Eisenhower matrix to help agents distinguish between urgent and important work.
  • Development relies on modern stacks including Next.js 16, React 19, and Bun workspaces.

The Bottom Line

This is a clever move to ground AI agents in real-world productivity logic. By exposing a structured matrix via MCP, GSD gives agents the ability to prioritize rather than just execute, which is the missing link in many agentic workflows.