add-gmail

Integrate Gmail with NanoClaw for reading, sending, and email-triggered workflows.

Updated Feb 24, 2026
One-click install
npx skills add https://github.com/webboty/flexclaw --skill add-gmail-webboty
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: add-gmail
Source: https://github.com/webboty/flexclaw/tree/main/.claude/skills/add-gmail
Command: npx skills add https://github.com/webboty/flexclaw --skill add-gmail-webboty

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill simplifies adding Gmail access to NanoClaw so agents can read, send, and be triggered by email without exposing host credentials or manual container edits. It removes the friction of GCP OAuth setup, container mounting, and agent tooling so email workflows integrate reliably with existing agent runners.

Core Features & Use Cases

  • Tool Mode: Allow agents to read and send messages only when invoked from another channel (for example, WhatsApp-triggered actions).
  • Channel Mode: Enable email-based triggers, scheduled tasks, and threaded replies by polling Gmail and routing messages into agent contexts.
  • Implementation Steps: Guides through creating GCP OAuth credentials, saving them to ~/.gmail-mcp, adding the Gmail MCP tool to the agent-runner, mounting credentials into containers, adding DB tracking for processed messages, and creating an email polling loop.
  • Use Case: Receive a support request by email, have NanoClaw process it in a dedicated conversation context, and send a threaded reply automatically.

Quick Start

Enable Gmail integration by running the OAuth setup, placing your GCP credentials in ~/.gmail-mcp, and choosing Tool or Channel mode so NanoClaw can read and send emails.

Frequently Asked Questions about add-gmail

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

FAQPage Schema
How do I integrate Gmail with an AI agent to trigger workflows from incoming emails?

You can integrate Gmail with an AI agent by setting up GCP OAuth credentials, mounting them into the agent container, and running an email polling loop. This allows incoming emails to trigger scheduled tasks and automated threaded replies.

What is the difference between using Gmail as a tool versus a channel for AI agents?

Using Gmail as a tool allows agents to read and send messages only when invoked from another channel like WhatsApp. Using Gmail as a channel enables email-based triggers, scheduled tasks, and threaded replies by actively polling the inbox.

How do I securely mount GCP OAuth credentials for an AI agent running in Docker containers?

To securely mount GCP OAuth credentials, save them to a local directory such as ~/.gmail-mcp, then mount that directory into your agent containers. This prevents exposing host credentials while allowing the agent to access Gmail.

Can I use WhatsApp to trigger an AI agent to read and respond to Gmail messages?

Yes, you can trigger Gmail actions from WhatsApp by using Tool Mode. This configuration allows agents to read and send emails only when invoked from a messaging channel, routing the context from WhatsApp to Gmail.

How do I prevent an AI agent from processing the same Gmail message twice?

You can prevent duplicate processing by adding database tracking for processed messages. The integration records each message in the database as it is polled, ensuring the agent only handles new emails.

Do I need to manually edit containers to give my agent access to Gmail?

No, you do not need to manually edit containers. The integration handles GCP OAuth setup, mounts the credentials directory into agent containers, and registers the Gmail MCP tool automatically.