email-guard-mcp

Email safety for AI agents: scan inbound for prompt-injection/phishing + outbound for secret leaks.