Skip to main content
Glama
mail-mcp-server

mail-mcp-server

create_draft

Compose new messages, replies, reply-alls, or forwards as unsent drafts in a connected mailbox so you can review them before planning to send.

Instructions

Write a draft: a new message, a reply (reply_to, optionally reply_all) or a forward. A draft never sends, so this runs directly; plan_send sends it.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "mailbox": "<mailbox>",
      "body": "<body>"
    },
    "name": "create_draft"
  }
}

Input Schema

NameRequiredDescriptionDefault
ccNo
toNoRecipients; for a reply, defaults to the sender being answered
bccNo
bodyYesPlain-text body
fromNoA send-as address of the mailbox; default is its default identity
forwardNoMessage id to forward, with its text and attachments
mailboxYesA connected mailbox address (see list_mailboxes)
subjectNoRequired for a new message; a reply or forward derives Re:/Fwd:
reply_toNoMessage id being replied to (keeps the thread)
reply_allNoWith reply_to: also copy the other recipients

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.2.0

TDQS

A4.3/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations supplied, the description carries the full burden and does disclose the decisive behavioral trait: this executes directly rather than producing a plan, and has no send side effect. It leaves unstated the return value (presumably a draft id needed to later plan_send) and any per-mailbox authorization requirements.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences, front-loaded with the verb and the mode enumeration, then the critical non-sending constraint and the alternative. No filler or restatement of the tool name.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 10-parameter tool with no annotations and no output schema, the description covers mode selection and the no-send behavior well, but says nothing about what is returned or how attachments are carried through a forward. That leaves a meaningful gap given the absence of an output schema to compensate.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Beyond the 80%-covered schema, the description explains how parameters combine into modes: reply_to drives reply behavior, reply_all is conditional on it, and forward is the third path. That mode-selection semantics is real added meaning the field-level descriptions alone don't convey.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ("Write a draft") and then enumerates the three modes it supports: new message, reply (with reply_to/reply_all), forward. It also names the sibling boundary explicitly ("plan_send sends it"), so an agent can tell it apart without opening another schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives clear selecting context: use this when you only want to compose, because "a draft never sends"; use plan_send when you want it sent. It does not address editing an existing draft or attachment handling, but the core when/when-not distinction is stated plainly.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.