跳到正文
cccskills
免费在 GitHub 公开

agentmail

Inter-agent communication for tmux sessions. Send and receive messages between AI agents.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(2)

  • SKILL.md3.5 KB
  • metadata.json1.3 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

AgentMail Skill

AgentMail enables communication between AI agents running in different tmux windows through a simple file-based mail system.

Overview

AgentMail is a CLI tool for inter-agent communication within tmux sessions. Messages are stored in .agentmail/mailboxes/ as JSONL files, providing persistent, file-locked message queues for each agent.

Prerequisites

  • Must be running inside a tmux session
  • AgentMail CLI must be installed and available in PATH
  • Messages stored in .agentmail/ directory

Core Commands

Send a Message

agentmail send <recipient> "<message>"

Send a message to another agent. The recipient must be a valid tmux window name.

Examples:

agentmail send agent2 "Can you review the changes in src/api?"
agentmail send -r worker -m "Task completed"
echo "Build succeeded" | agentmail send agent2

Receive Messages

agentmail receive

Read the oldest unread message from your mailbox. Messages are delivered in FIFO order and marked as read after display.

List Recipients

agentmail recipients

List all tmux windows that can receive messages. Your current window is marked with [you].

Set Status

agentmail status <ready|work|offline>

Set your availability status:

  • ready - Available to receive messages and notifications
  • work - Busy working (suppresses notifications)
  • offline - Not available (suppresses notifications)

Workflow

Checking for Messages

  1. Run agentmail receive to check for new messages
  2. If a message is available, read and process it
  3. Optionally reply using agentmail send

Sending Messages

  1. Run agentmail recipients to see available agents
  2. Send your message: agentmail send <recipient> "<message>"
  3. Confirm delivery via the returned message ID

Status Management

The plugin automatically manages your status:

  • Session start: Status set to ready
  • Session end: Status set to offline
  • End of turn (Stop): Status set to ready, checks for new messages

Message Format

Messages include:

  • ID: Unique 8-character base62 identifier (a-z, A-Z, 0-9)
  • From: Sender's tmux window name
  • To: Recipient's tmux window name
  • Content: Message body

Best Practices

  1. Check messages regularly - Use agentmail receive to stay informed
  2. Keep messages concise - Focus on actionable information
  3. Include context - Reference files, line numbers, or specific details
  4. Respond promptly - Other agents may be waiting for your input
  5. Use status appropriately - Set work when focusing on complex tasks

Integration with Claude Code

This plugin integrates with Claude Code hooks:

  • SessionStart: Automatically sets status to ready and runs onboarding
  • SessionEnd: Automatically sets status to offline
  • Stop: Sets status to ready and checks for new messages

The hooks ensure agents are properly registered and can receive notifications from the mailman daemon.

Troubleshooting

"Not in a tmux session"

AgentMail requires tmux. Start a tmux session first.

"Recipient not found"

The recipient window doesn't exist. Check available windows with agentmail recipients.

"No unread messages"

Your mailbox is empty. Other agents haven't sent you messages yet.

Messages not being delivered

Ensure the mailman daemon is running: agentmail mailman --daemon

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

[00] TnR ABSOLUTE RIGOR — Systemic Verification Doctrine. Forced cognitive exertion for high-stakes investigations. Mandates absolute deconstruction, mathematical proof chains, and ≥95% derived certainty. Zero-tolerance for heuristics or sentiment-based reasoning. Shortcuts are defined as intentional hallucinations.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

2agent

免费

Configures 2-Agent workflow between PM and implementation roles. Use when user mentions 2-Agent, 2エージェント, PM連携設定, Cursor設定, Cursor連携, 2-agent運用. Do NOT load for: 単独運用, ワークフロー実行, ハンドオフ処理.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

Appends relevant hashtags to journal entries based on content analysis. Intended for LLM execution.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

add-agent

免费

Add a new AI coding agent to Agent Sandbox. Creates all required files (Dockerfile, templates, CI, docs) and wires the agent into the CLI, proxy, and build system.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

Add a new property to the AI agents database. Use when the user wants to add, create, or introduce a new column, property, field, or feature to track across all agents in the comparison matrix. Handles all four required steps - database updates, groups.json, table display, and GitHub issue templates.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

Add new Agent templates to the meta-agent-skills framework.

中文简介正在准备中。当前显示原文说明。

diegosouzapw/awesome-omni-skill642026年3月2日 更新

查看 diegosouzapw 的全部技能

报告此技能的问题