$ loading_
Enable AI agents to send physical mail and receive scanned inbound mail.
The materials indicate this MCP tool integrates with mailbox.bot mail services and is open-source under MIT, but the available information is very limited. No explicit remote endpoints or required secrets are listed, so the overall posture is closer to low-to-moderate risk; however, its stated features involve physical mail, scanned inbound mail, and webhooks, so network behavior and data handling should still be verified before deployment.
The provided materials explicitly state that no keys or environment variables are required. No API key, token, or other sensitive credential is documented, so there is no clear credential exposure or abuse surface based on the known facts; however, the description mentions sandbox keys, and the missing documentation means the actual implementation should still be checked for implicit auth settings.
As an MCP server for mailbox.bot, its stated features include sending mail, receiving scanned inbound mail, and webhooks, which strongly suggests communication with the service. Although the checks say 'remote endpoint host: none,' this appears more like missing documentation than proof of no networking; if it actually sends mail content, addresses, or scanned documents to the service, that should be treated as ordinary egress and verified for target domains and data scope before use.
The system marks this tool as executes-code, meaning it runs locally as an MCP service and executes code. This is a normal capability for MCP tools, and the current materials do not show requests for unusual system privileges or unrelated high-risk actions, so this is rated as caution rather than risk.
The description indicates it handles potentially sensitive business data such as mail content, delivery details, and scanned inbound mail. The current materials do not specify which local files it reads or writes, whether it persists data, or how approval controls are implemented, so the data access surface cannot be fully audited; based on its function, it should be assumed to handle user-provided letter content, recipient addresses, and scanned documents.
Positive factors include being open-source under an MIT license, making the source in principle auditable. The concerns are that it comes from a third-party registry, has 0 stars, unknown maintenance status, and no README, leaving limited audit context; this suggests insufficient supply-chain transparency rather than clear maliciousness, so it is rated as caution.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "mailbox-mcp" yet — see the docs or source repo.
Use mailbox-mcp to create a certified mail letter to Zhang San at 88 Example Road, Pudong, Shanghai. The subject is contract signature reminder. Write a formal, concise body and return tracking details and approval status.
A certified letter draft, sending request result, approval status, and trackable mail details.
Use mailbox-mcp to batch-create thank-you postcards for the following customer list, including an appreciation message and brand signature. If human approval is required, summarize all pending approval items first. List: ...
A bulk mailing job with per-recipient status and a summary of pending approvals.
Use mailbox-mcp to check for new scanned inbound physical mail. If any exists, extract the sender, date, and key content, then organize them into an action summary.
New scanned mail records, extracted key details, and an actionable summary.
Create reviewable physical mail drafts, verify quotes, and pay postage through AI agents.
Connect AI agents to email for sending, organizing, and workflow automation.
Enable AI agents to securely send emails with SMTP and attachments.
Manage multi-agent messaging, task delegation, resource coordination, and monitoring.
Enable AI agents to send, receive, and manage email workflows programmatically.
Enable AI assistants to send SMTP emails with attachments and connection testing.