독립형 Gmail MCP Server — 초안 전용입니다. Lane은 절대 전송하지 않으며, 단일 도구가 소유자가 검토 후 직접 전송할 Gmail 초안을 생성합니다.
gmail_draft_composeGmail 초안을 생성합니다.
인수 | 유형 | 필수 | 참고 |
| string | 아니요 |
|
| string | 아니요 | 알 수 없으면 비워 두세요 — 소유자가 초안을 검토할 때 채웁니다. |
| string | 예 | |
| string | 예 | 일반 텍스트 |
{ draftId, messageId, draftUrl }을 반환합니다.
이 서버에는 전송 도구, 읽기 도구, 라벨/스레드 관리 기능이 없습니다 — 초안만 작성합니다.
Related MCP server: Gmail MCP Server
npm install
npm run dev # run the server via tsx
npm run build # compile to dist/
npm run check # test + buildgmail.compose 범위만 있는 OAuth 2.0 액세스 토큰이 필요합니다 — gmail.send는 절대 사용하지 않습니다. GMAIL_ACCESS_TOKEN으로 설정하거나 호출할 때마다 accessToken을 전달하세요. .env.example을 참조하세요.
Create a Gmail draft. Draft-only — this server has no send capability. The owner reviews the draft in Gmail and sends it themselves.
| Name | Required | Description | Default |
|---|---|---|---|
| to | No | Recipient email address. Leave empty if unknown — the owner fills it in when reviewing the draft. | |
| body | Yes | Email body, plain text | |
| subject | Yes | Email subject line | |
| accessToken | No | Gmail OAuth 2.0 access token. Falls back to GMAIL_ACCESS_TOKEN env var. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden of behavioral disclosure. It explicitly reveals the key behavioral trait — that drafts are created but never sent — which prevents the agent from assuming a full send pipeline exists. This goes beyond what a generic 'create draft' description would provide, though it omits details like draft location or return value.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences with no filler. The core action is front-loaded, and the critical caveat (draft-only, no send capability) is stated immediately after in the same breath. Every word earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple 4-parameter tool with no enums, nested objects, or siblings, the description plus fully-documented schema cover the essential operating context. The draft-vs-send distinction is the most important edge case for an agent to know, and it is explicit. An output schema is absent, but the description compensates reasonably well.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3 even though the tool description itself adds no parameter-level detail. The schema already documents each parameter sufficiently, including the nuance that 'to' may be left empty for the owner to fill in. The description does not need to repeat or expand on them.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
"Create a Gmail draft" states a specific verb plus an unambiguous resource, and immediately clarifies the draft-only scope. The clarification that the server has no send capability actively prevents an agent from mistakenly believing this tool can deliver email.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description establishes a clear workflow: the agent creates the draft, and the owner reviews and sends it in Gmail. However, there are no explicit use-this-when or when-not conditions, and with no sibling tools listed, the guidance is contextual rather than directive.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.0gmail_draft_composeScored across 1 tool
Only one tool exists, so there is no possibility of confusion between tools. The single tool's purpose is clearly stated.
The single tool name follows a clear verb_noun pattern (gmail_draft_compose), but with only one tool there is limited evidence of a broader convention. It is readable and predictable.
A single tool for a Gmail server is far too thin to be useful. The server's apparent scope is much broader than drafting, and one tool cannot reasonably cover it.
The surface is severely incomplete. There is no send, read, list, search, or delete capability, and the tool only creates drafts, leaving the agent with a dead end after composing.
Manage Gmail end-to-end: search, read, send, draft, label, and organize threads. Automate workflow…
Manage Gmail messages, threads, labels, drafts, and settings from your workflows. Send and organiz…
Email inboxes for AI agents: send, reply, read verification codes, human-approved drafts.
Email inboxes and calendars for AI agents: send, receive, search, draft and schedule.