MCP Server

Active

Apache-2.0

ai.packmind/mcp-server MCP Server

ai.packmind/mcp-server

Sync one engineering playbook of coding standards and skills to every AI coding agent via MCP.

What is the ai.packmind/mcp-server MCP server?

Packmind MCP server is a remote MCP endpoint that connects AI coding agents (GitHub Copilot, Claude Code, Cursor, etc.) to Packmind, a platform for capturing, centralizing, and distributing an organization's engineering standards, commands, and skills. It lets agents create, retrieve, and manage coding standards directly from within the AI tool, so teams don't have to manually maintain separate instruction files (CLAUDE.md, .cursor/rules, copilot-instructions.md, AGENTS.md) per repo and per agent.

Packmind centralizes an organization's technical decisions—architecture rules, naming conventions, patterns, best practices—into a single engineering playbook, then distributes it as the exact instruction files each AI coding tool needs (Copilot, Claude Code, Cursor, AGENTS.md, etc.). The MCP server exposes this playbook directly to AI agents, allowing them to create and manage standards and commands interactively from within the coding session, instead of requiring engineers to copy-paste and keep multiple instruction file formats in sync across repos. It differs from a plain shared Git repo/marketplace by adding an ownership/approval workflow, in-codebase update proposals, multi-agent rendering from one source, and adoption tracking.

How to install ai.packmind/mcp-server

Copy-paste configuration for popular MCP clients.

transport: http

Config generated by PluginBench — verify against the source before use.

Environment / auth

Authorization
Bearer token for Packmind authentication

Cursor

~/.cursor/mcp.json

{
  "mcpServers": {
    "mcp-server": {
      "url": "https://app.packmind.ai/mcp"
    }
  }
}

Windsurf

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "mcp-server": {
      "serverUrl": "https://app.packmind.ai/mcp"
    }
  }
}

VS Code

.vscode/mcp.json

{
  "servers": {
    "mcp-server": {
      "type": "http",
      "url": "https://app.packmind.ai/mcp"
    }
  }
}

Claude Desktop

Claude Desktop connects to remote servers via the mcp-remote bridge.

{
  "mcpServers": {
    "mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://app.packmind.ai/mcp"
      ]
    }
  }
}

Claude Code

claude mcp add --transport http mcp-server https://app.packmind.ai/mcp

Tools & capabilities

Tools this server exposes to the agent.

Start packmind onboarding — An interactive prompt run in the AI agent that guides the user through creating their first coding standard from their codebase.

Use cases

Onboard a new AI coding agent to an organization's existing engineering standards via a guided interactive prompt
Create new coding standards, commands, or skills directly from within an AI coding session based on the current codebase
Keep coding standards synchronized and rendered correctly across Copilot, Claude Code, Cursor, and AGENTS.md without maintaining separate files
Submit and manage updates to team standards from within a project rather than editing a separate central repo
Track which context files/standards are adopted, in which repos, and at which version

ai.packmind/mcp-server MCP server FAQ

What does the Packmind MCP server do?

It connects your AI coding agent (Copilot, Claude Code, Cursor, etc.) to Packmind so the agent can create, retrieve, and manage your organization's coding standards, commands, and skills directly during a coding session, and keep them synced across every tool and repo.

Is the Packmind MCP server free?

Packmind offers a free cloud account at app.packmind.ai, as well as a self-hosted option deployable via Docker Compose or Kubernetes for organizations that want to run it on their own infrastructure.

How do I connect the Packmind MCP server in Cursor or Claude Code?

In Packmind, go to Account Settings and copy your MCP Access token, then configure your AI agent with the MCP server URL ({PACKMIND_URL}/mcp, e.g. https://app.packmind.ai/mcp) and the access token. After setup, run 'Start packmind onboarding' in the agent.

What authentication is required?

A Packmind MCP Access token, generated from the Account Settings page in your Packmind organization, is required to authenticate the AI agent against the remote MCP server.

How is this different from just using a shared Git repo of prompts/rules?

Packmind adds controlled editing with an ownership/approval workflow, lets updates be proposed from the project codebase itself, renders one source into the specific formats each AI tool needs (CLAUDE.md, .cursor/rules, copilot-instructions.md, etc.), and tracks adoption across repos and versions.

Can I use Packmind without the MCP server?

Yes, Packmind also offers a CLI (packmind-cli init) as an alternative setup option, which works alongside agent-specific commands like /packmind-onboard.

README (reference)

Source of truth, from the repository.

One Engineering Playbook. Synced Everywhere. For Every AI Coding Agent.

License Stars

❗ The 2 big problems every AI-native engineer runs into

1️⃣ “What do I even put in these AI instructions?”

Every tool expects its own inputs:

Copilot → .github/copilot-instructions.md, chat modes, reusable prompts
Claude → CLAUDE.md, commands, skills
Cursor → .cursor/rules/*.mdc, commands, skills
AGENTS.md → AGENTS.md
(with more formats appearing every month…)

But your team’s actual standards aren’t stored anywhere:

architecture rules → buried in Slack or Notion
naming conventions → stuck in your head
patterns → hiding in PR comments
best practices → scattered across repos

👉 Packmind helps you turn all of this into a real engineering playbook (standards, commands, skills) so AI agents finally code your way.

2️⃣ “Why am I copy-pasting this across every repo and every agent?”

Every repo. Every assistant. Different files, different folders, different formats.

Keeping everything in sync is impossible.

👉 Packmind centralizes your playbook once — and distributes it everywhere, generating the exact instruction files each AI tool needs, optimized for context.

🆚 Why Packmind over a plain Claude Code marketplace or a plain centralized Git repository?

A marketplace distributes skills and commands from a Git repo. Packmind does more:

Controlled editing: context files go through a clear ownership and approval workflow. No PR discipline or CODEOWNERS conventions to enforce.
Simplified updates: update proposals are submitted from the project codebase, no separate repo to clone or PR.
Multi-agent: one source, rendered for Claude Code, Copilot, Cursor and more. No parallel CLAUDE.md or .cursor/rules to maintain.
Adoption tracking: see which context files are used, in which repo, at which version.

A marketplace ships content. Packmind governs it.

Get started

Choose your preferred setup option:

Cloud version: Get started at https://app.packmind.ai (free account)
Self-hosted: Deploy on your own infrastructure using Docker Compose or Kubernetes

Option 1: Install the CLI (recommended)

Follow the instructions during the onboarding to connect to your Packmind organization You can find them at anytime in the Settings menu.

Once authenticated, run in your project:

$> packmind-cli init

Then, in your favorite ai coding agent, run:

/packmind-onboard

To create your first standards and commands from your codebase.

Option 2: Connect MCP server

The MCP server allows you to create and manage standards and commands directly from your AI agent (GitHub Copilot, Claude Code, Cursor, etc.).

Go to Account Settings in Packmind
Copy your MCP Access token
Configure your AI agent with:
- MCP server URL: {PACKMIND_URL}/mcp
- Your MCP access token

Once set up, open your AI agent and use this prompt:

Start packmind onboarding

Your AI agent will guide you through creating your first coding standard interactively.

Documentation

Available here: https://docs.packmind.com.

:compass: Key Links

View repository →