agent-mail

Coordinate multiple coding agents via a Git-backed MCP Agent Mail server.

4|2|Updated Sep 20, 2023
One-click install
npx skills add https://github.com/javdl/nixos-config --skill agent-mail-javdl
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-mail
Source: https://github.com/javdl/nixos-config/tree/main/mcp_agent_mail
Command: npx skills add https://github.com/javdl/nixos-config --skill agent-mail-javdl

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires fastmcp, GitPython, pathspec, markdown2, bleach, sqlmodel, sqlalchemy, aiosqlite, httpx, python-decouple, and includes scripts (resource) components.

What problem does it solve?

Coordinating multiple coding agents across one or more codebases can lead to conflicts, context loss, and chaotic messaging. MCP Agent Mail provides a Git-backed, auditable inbox/outbox with persistent identities, advisory file reservations, and a Human Overseer workflow to keep parallel work safe and transparent.

Core Features & Use Cases

  • Memorable agent identities (e.g., GreenCastle) bound to project context.
  • Threaded messaging with per-recipient inboxes and Markdown-rich bodies.
  • Advisory file reservations (leases) to signal intent before edits.
  • A pre-commit guard that blocks conflicting commits against active reservations.
  • Human Overseer for urgent guidance and cross-team coordination.
  • Git+SQLite persistence for auditable history and fast search; supports static export for offline review.
  • Disaster recovery tooling and static export pipelines to share mailboxes.

Quick Start

Start the MCP Agent Mail server and browse the UI at /mail to begin coordinating agents.

Frequently Asked Questions about agent-mail

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I coordinate multiple coding agents to prevent file edit conflicts?

You can coordinate multiple coding agents by using advisory file reservations, or leases, that signal intent before edits. A pre-commit guard then blocks conflicting commits against active reservations to keep parallel work safe.

How does MCP agent coordination work with git-backed messaging?

MCP agent coordination works through a centralized server using git-backed messages and threads. It provides per-recipient inboxes with Markdown-rich bodies and uses SQLite for auditable history and fast search.

Can I use this agent coordination workflow across multiple git repositories?

Yes, this workflow applies across both single-repo and multi-repo work setups. It enables cross-project coordination with human oversight using memorable agent identities bound to project context.

Do I need Python and FastMCP to run an asynchronous agent mail server?

Yes, running this asynchronous agent mail server requires Python, FastMCP, Git, and SQLite. These dependencies provide the Git-backed, auditable inbox and outbox infrastructure for persistent agent identities.

What is the best way to oversee urgent guidance for parallel coding agents?

The best way to oversee urgent guidance is through a Human Overseer workflow. This workflow enforces safe collaboration and handles urgent guidance and cross-team coordination for parallel agents.

How do I export agent mailboxes for offline review if a disaster occurs?

You can export agent mailboxes for offline review using static export pipelines. The system also provides disaster recovery tooling to help manage and restore your auditable coordination history.