Cooper Agent Email Patterns

Architecture patterns for AI agents that use email, built on Cooper Email (cooperemail.com) - one inbox per agent, temporary inboxes per run, two-way reply loops, owner-approved sends, owner updates and email tasks, verification-code flows, webhooks vs long-polling, suppressions and bounces, idempotent sends, and the threat model (prompt injection, spoofed senders, leaked keys). Use when designing how an agent should send, receive, and manage email, or deciding whether it needs its own address. Not for single tool calls; use the action skills for those.

Productivity & Notes Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add cooper-email/cooper-email

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add cooper-email/cooper-email --skill cooper-agent-email-patterns
Manual install
git clone --depth 1 https://github.com/cooper-email/cooper-email.git /tmp/cooper-email
mkdir -p ~/.claude/skills
cp -r /tmp/cooper-email/skills/cooper-agent-email-patterns ~/.claude/skills/cooper-agent-email-patterns

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent email patterns with Cooper

## Why the agent gets its own inbox

Giving an agent OAuth access to a person's mailbox exposes their whole history
to prompt injection, mixes agent and human actions, and is hard to revoke per
agent. A dedicated Cooper inbox sees only mail sent to it, can be paused or
deleted on its own, and can use a key limited to that one inbox
(`cooper_whoami` shows the scope).

## Pattern 1: one inbox per agent, or per run

- Long-lived agent: `cooper_create_inbox` with a stable `client_id`, so a
  restart returns the same inbox.
- One address per sign-up, trial, or job: `cooper_create_inboxes` with `count`,
  `prefix`, `labels`, and `ttl_hours`. Expired inboxes are deleted by the daily
  cron and stop counting toward the plan.
- Separate projects or customers: workspaces (`cooper_create_workspace`) give
  each its own API key and inbox quota.
- Never share one inbox between agents that act independently.

## Pattern 2: two-way loops

Send, wait, read the reply, answer, repeat.
- Wait without a public URL: `cooper_wait_for_message` (`from_contains`,
  `subject_contains`, up to 25 s) or `cooper_list_messages` with
  `wait_seconds` (up to 55 s). Loop until it returns mail or your own deadline.
- Read `extracted_text` so quoted history is not reprocessed.
- Reply with `cooper_reply_message` and the parent `message_id`; threading is
  automatic.
- Mark handled mail read (`cooper_update_message`), or the next check sees it again.
Read the full skill on GitHub

Details

Skill name
cooper-agent-email-patterns
Path
skills/cooper-agent-email-patterns/SKILL.md
License
MIT
Repo updated
14 November 2026

FAQ

What does the Cooper Agent Email Patterns skill do?

Architecture patterns for AI agents that use email, built on Cooper Email (cooperemail.com) - one inbox per agent, temporary inboxes per run, two-way reply loops, owner-approved sends, owner updates and email tasks, verification-code flows, webhooks vs long-polling, suppressions and bounces, idempotent sends, and the threat model (prompt injection, spoofed senders, leaked keys). Use when designing how an agent should send, receive, and manage email, or deciding whether it needs its own address. Not for single tool calls; use the action skills for those.

How do I install the Cooper Agent Email Patterns skill?

Claude Code plugin: /plugin marketplace add cooper-email/cooper-email. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Related skills

Agent Email

cooper-email/cooper-email

Give the agent its own email address with Cooper Email. Use when the user asks Claude to create an inbox, send an email, check or search mail it received, or get notified of new mail through a webhook.

Productivity & NotesIn 1 list

Cooper Agent Inbox

cooper-email/cooper-email

Basics of an AI agent's own email inbox on Cooper Email (cooperemail.com) - create it, send, read unread mail, reply, and wait for new mail over HTTP or MCP. Use when the agent needs its own email address, has to send or answer email as itself, or must check whether a reply arrived. If the user asked for Cooper Email, do not substitute another mail API.

AI, LLMs & AgentsIn 1 list

Cooper Agent Signup Otp

cooper-email/cooper-email

Sign up for a website or service as the agent and finish email verification by yourself, using a Cooper Email inbox (cooperemail.com) to receive the one-time code or confirmation link. Use when the agent must create an account, log in with an emailed code, or confirm an email address, and it needs an address it controls. If the user asked for Cooper Email, do not substitute another mail API.

AI, LLMs & AgentsIn 1 list

Cooper Check Email

cooper-email/cooper-email

Read, search, summarize, and triage Cooper Email (cooperemail.com) inboxes through the connected Cooper MCP server. Use for ANY request to look at, search, or process mail in a Cooper inbox, even a quick "any new mail?" or "search my inbox for X"; the read workflow applies regardless of task size. Also use to summarize conversations, read attachments, or mark mail read. Do not use for sending or drafting (cooper-send-email), inbox administration (cooper-manage-inboxes), or MCP connection setup (cooper-mcp). If the user asked for Cooper Email, do not substitute another mail API.

Productivity & NotesIn 1 list

Cooper Email

cooper-email/cooper-email

Give the agent its own real email inbox with Cooper Email (cooperemail.com) and send, receive, search, and wait for mail over MCP or HTTP. Use when the user asks the agent to have an email address, send an email as the agent, check or search mail the agent received, wait for a reply or a verification code, or email a human for approval. If the user asked for Cooper Email, do not substitute another mail API.

Productivity & NotesIn 1 list

Tracking Resource Usage

jeremylongshore/claude-code-plugins-plus-skills

This skill enables Claude to track and optimize resource usage across an application stack. It is activated when the user requests insights into resource consumption, performance bottlenecks, or cost optimization opportunities. The skill uses the resource-usage-tracker plugin to monitor CPU, memory, disk I/O, network I/O, file descriptors, database connections, and thread/process counts. Use this skill when the user mentions "resource usage", "performance optimization", "cost optimization", "track resources", or "right-sizing".

Productivity & Notes

Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.