MCP server · Developer tools
Agentpack MCP server
by ihorponom
Helps your AI coding assistant remember task state between sessions so it can pick up where it left off.

Agentpack is a small helper that keeps a task notebook inside your code project. It is made for people who use AI coding assistants like Claude Code, Cursor, or Codex and get tired of the assistant forgetting everything when the chat restarts. If you often come back to the same bug or feature days later, this is for you.
What is an MCP server? The 30-second version
On its own, your AI can only chat with you. An MCP server is a small helper program that gives your AI a new skill or a connection to something useful. This one connects your AI to a task ledger stored in your project folder, so the AI can read what was decided before and write down what it learns now. That way the next session, even in a different app, can continue the same task.
What this MCP server does
You ask your AI to work on a task, like fixing a flaky test. The AI uses this helper to load the current task state from a folder called .agentpack in your project. While working, it records decisions, dead ends, and evidence through the helper. When you stop, it saves a checkpoint with the status and next steps. Next time you open any connected AI client, it loads that state and continues without re-investigating.
Click to zoomWhat you can do with it
- Load the current task state at the start of a session
- Record decisions and approaches that did not work
- Save checkpoints with status, next actions, and git state
- Cache conclusions about files with a hash so they are not re-read
- Warn when edits happen outside the active task's folder
- Inspect the same task state yourself from the command line
- Hand off a task to another agent or a teammate's agent
Try asking your AI
- “Load the Agentpack task state for the flaky test fix and continue where we stopped”
- “Record that we tried retrying the API call and it did not help”
- “Create a checkpoint for the current task with next actions”
- “Show me the current Task Passport and recent checkpoints”
What it gives back to you
You get a short, compressed summary of the task state: the goal, status, recent decisions, dead ends, and next actions. It also shows checkpoints and which files were reviewed. In the chat, this looks like a plain text summary the AI reads before continuing. When the AI records something, it confirms what was saved.
Before you start
What you need
- Node.js version 20 or newer
- A code project folder on your computer
- An AI coding client that supports MCP, like Claude Code, Cursor, or Codex
- The agentpack-cli tool installed (npm install -g agentpack-cli)
Good to know
The task gate can warn or block edits outside the active task folder, and the ledger stores notes about your project in a .agentpack folder, so review what gets recorded if your work is sensitive.
Install it with your AI
Add Agentpack MCP server to your AI, no technical skills needed
You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.
Sign in to get the install prompt
Members get a ready-made prompt that lets the Claude desktop app check Agentpack MCP server, install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.
Who it's for
Developers and anyone who uses AI coding assistants on long or multi-day tasks and wants the assistant to remember what happened.





