Atlas / MCP servers / yunfeizhu / Mail

MailSAFE

mcp/yunfeizhu/mail-1

A lightweight Model Context Protocol (MCP) server that provides IMAP and SMTP email functionality for AI assistants like Claude in Cursor IDE. Built with TypeScript and optimized for easy deployment via npm/npx.

Verdict
SAFE
Grade
B
Trust score
89 /100
Exposed tools
12 7r · 4w · 1d
Transport
stdio
License
MIT
Stars
51
01

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Give your AI assistant a local bridge to your email.

Search, read, organize, reply to, and send email through supported password- or app-password-based IMAP/SMTP accounts—from Claude, Cursor, Codex, and other MCP clients.

[](https://www.npmjs.com/package/mcp-mail-server) [](https://www.npmjs.com/package/mcp-mail-server) [](https://www.npmjs.com/package/mcp-mail-server) [](LICENSE)

English · 简体中文

Why this server · Quick start · Client setup · Tools · Configuration · Development

“Find the unread messages from Alice this week, summarize them, and move the finished thread to Archive.”

Why this server?

🔎 Find what mattersSearch across folders by sender, recipient, subject, body, date, read state, and reply state. ✉️ Act without leaving the conversationRead, send, reply, move, and delete messages using natural language.

📎 Work with attachmentsInspect metadata, download files, and send local attachments through explicit filesystem allowlists. 🔐 Keep controlRun locally, connect directly to your mail provider, verify TLS certificates, and enforce payload limits.

What can I ask?

Read from source at commit cd2280b9ee4bOBSERVED · 2026-10-08
02

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code
claude mcp add mcp-mail-server -- npx -y [email protected]
claude-desktop
{
  "mcpServers": {
    "mcp-mail-server": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ]
    }
  }
}
03

Exposed tools (12)

7 read · 4 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

ToolRiskDescription
check_connectionreadConnect to IMAP when needed, actively verify SMTP, then return structured status for both services.
continue_email_threadreadFind the newest sent message with the same normalized subject, optionally narrowed by recipient and date, then continue its thread. Designed for recurring reports; reply-all and quoted history both default to true.
delete_messagedestructivePermanently delete one existing mailbox-scoped message using targeted UID EXPUNGE. Fails safely when the UID is missing or the IMAP server lacks UIDPLUS, and returns a structured unknown outcome when a transport failure prevents confirmation.
find_unreplied_messagesreadFind messages from a sender that have no thread-header reply in the sent mailbox. Messages without Message-ID are returned separately as unknown.
get_messagewriteRetrieve one message, including full bodies and attachment metadata. Set markSeen only when the message should be marked as read.
get_messagesreadRetrieve multiple messages from one mailbox, including full bodies and attachment metadata. Fails if any requested UID is missing or the aggregate response-body budget would be exceeded.
list_mailboxesreadList all available mailboxes (folders). Auto-connects to IMAP when needed.
move_messagewriteMove an existing mailbox-scoped message to another existing mailbox. If a move, copy, or source-cleanup outcome cannot be confirmed, returns a structured partial error that must be inspected before retrying.
reply_to_emailreadReply to a mailbox-scoped message with threading headers, optional reply-all recipients, signature, and quoted original content.
save_attachmentwriteSave one or all attachments from a mailbox-scoped message into an allowed local directory. Select one attachment by zero-based index or exact filename. If a later write fails, returns the files already saved as a structured partial result.
search_messagesreadSearch one or more mailboxes with combined filters. Defaults to INBOX and the detected sent mailbox. Returns message summaries unless includeBody is true; full-body hydration fails explicitly if a selected message disappears or the aggregate response-body budget would be exceeded.
send_emailwriteSend a new email via SMTP and try to save a MIME-equivalent copy with the accepted Message-ID to the sent mailbox.
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfaceWARN
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
declared (1 observation(s))
Network
none-observed
Shell
none-observed
Dependencies
not all pinned
Secrets in source
none-found

Findings (4)

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59
delete_message
Why it matters. 1 tool(s) can delete or overwrite
Fix. prefer a read-only mode or scoped tokens; the page states the blast radius
LOWInventory / provenance · inv.hidden_file · CWE-1104
.prettierignore
.prettierignore
Why it matters. hidden member outside the usual dotfiles
Fix. review its purpose
LOWInventory / provenance · inv.hidden_file · CWE-1104
.prettierrc.json
.prettierrc.json
Why it matters. hidden member outside the usual dotfiles
Fix. review its purpose
LOWSupply chain · supply.unpinned · CWE-829, CWE-1357
package.json
@modelcontextprotocol/sdk, imap, mailparser, nodemailer, zod, @eslint/js, @rollup/plugin-node-resolve, @rollup/plugin-terser
Why it matters. 20 dependency range(s) float
Fix. pin exact versions or ship a lockfile

Gates applied: no_behavioural_pass.

Audited 2026-10-08 · audit v0.4.1 · source sha cd2280b9ee4bfull audit observations/trust-audit/mcp-server/yunfeizhu__mail-1.json · Report an issue / request a re-scan
05

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

DateSourceVerdictGradeScoreChange
2026-10-08cd2280b9ee4bSAFEB89first audit
06

Questions

What is the Mail MCP server?

A lightweight Model Context Protocol (MCP) server that provides IMAP and SMTP email functionality for AI assistants like Claude in Cursor IDE. Built with TypeScript and optimized for easy deployment via npm/npx.

What tools does Mail expose?

12 in total: 7 read-only, 4 that write, and 1 that can delete or overwrite (delete_message). Every one is listed on this page with its risk.

Is Mail safe to connect to an agent?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.

What credentials does Mail need?

No credential environment variables were found in its source, so it appears to need none.

How does Mail run?

It speaks stdio, so it runs as a local process your client starts. It is published on npm as mcp-mail-server at 2.1.0.

How current is this page?

The grade is for one exact copy of the source (cd2280b9ee4b), read on 2026-10-08. The repository is watched and re-audited when it changes.

Advertisement