clawmail

What This Skill Does

ClawMail provides a robust, decentralized email infrastructure specifically engineered for AI agents and LLMs. Moving beyond traditional password-based authentication, ClawMail leverages Ethereum wallet signatures (EIP-191) to secure agent identities. By integrating directly into the Model Context Protocol (MCP), it enables seamless, tool-based email management, allowing agents to monitor, read, and process incoming communications autonomously. Whether you are using a free tier with a random address or paying a 1 USDC fee for a custom domain via Base mainnet, ClawMail ensures your agent has a permanent, authenticated mailbox to interface with the outside world.

Installation

There are two primary ways to integrate ClawMail. For Claude Desktop users, add the following configuration to your claude_desktop_config.json:

{ "mcpServers": { "clawmail": { "command": "npx", "args": ["clawmail"] } } }

Alternatively, for CLI usage, you can install the package globally via npm: npm install -g clawmail. If you are using the OpenClaw ecosystem, simply run: clawhub install openclaw/skills/skills/patrickshuff/clawmail-xyz.

Use Cases

ClawMail is ideal for agents that require external verification or communication workflows. It can be used for:

• Account Verification: Agents can sign up for services that require email confirmation and verify their own accounts by reading the confirmation emails.
• Automated Customer Support: Agents can act as the first line of support, reading incoming tickets sent to your custom address and responding or escalating based on message content.
• Subscription Monitoring: Use an agent to parse and summarize newsletters or system notifications sent to your agent's unique mailbox.

Example Prompts

1. "Check my clawmail inbox for any unread messages regarding my server deployment and read the most recent one for me."
2. "I need to verify my identity on a new service; please check if the email 'agent-bot' is available, then draft a plan to register it."
3. "List the last 10 messages in my mailbox and delete any that are promotional or unrelated to my current project tasks."

Tips & Limitations

• Security: Since authentication relies on wallet signatures, ensure your agent's wallet address is secure and only used for authorized operations.
• Limits: The free tier is restricted to 1000 messages; consider upgrading to the paid tier via USDC on the Base mainnet if your agent handles high-volume traffic.
• Privacy: Always review the headers of incoming messages, as agents may be susceptible to spam or phishing attempts just like human users.