sendgrid-mcp-secure
You can securely preview and send transactional emails through SendGrid, manage suppressions and templates, read email stats and domain auth, and inspect the server's security posture.
Send email safely via two-phase flow:
preview_emailvalidates/renders a payload and returns a single-use confirm token;send_emailsends only that exact payload. No direct-send path.Dry-run by default: in
dry-runmode writes are simulated/audited but not sent; switch toliveonly when ready.Manage suppressions:
add_suppressionstops future mail;remove_suppressionre-enables an address and requiresconfirm=true.Work with templates:
list_templatesandget_templatefetch dynamic/legacy templates and their versions.Get delivery stats:
get_email_statsreturns requests, delivered, opens, clicks, bounces, spam, aggregated by day/week/month.Check suppressions:
list_suppressionsandcheck_suppressioninspect all five suppression lists before drafting mail.Check domain authentication:
get_domain_authlists sending domains and DNS validity.See server posture:
server_statusshows mode, rate-limit remaining, allowlist, caps, audit-log path, and API-key presence (never the key).
Provides tools for sending transactional emails via SendGrid v3 API, including preview and two-phase send with confirm tokens, suppression list management, template retrieval, email stats, and domain authentication verification.
sendgrid-mcp-secure
A security-first MCP server for the SendGrid v3 API.
Email is an irreversible, reputation-bearing action. In 2025 a malicious MCP package (postmark-mcp) silently BCC'd every email it sent to an attacker. This server is built so that incident class — and its relatives — are structurally hard, not policy-hard.
Security model
Two-phase sends. There is no direct-send tool.
preview_emailvalidates and renders the exact payload, returns it for inspection, and mints a single-use confirm token (10-minute expiry).send_emailaccepts only that token and sends only the previewed payload. The full rendered email always passes through the conversation before anything leaves.Dry-run by default. Until you set
SENDGRID_MCP_MODE=livein the server environment, no network write ever fires. Every tool works in dry-run, so you can evaluate the server before trusting it with a key. (Read tools do still query SendGrid — dry-run stops writes, not traffic.)No BCC. Not accepted as an argument, off by default at the server layer. BCC injection was the postmark-mcp exfiltration channel.
Recipient allowlist (optional). Restrict sends to named addresses or
@domainsvia server env — immune to prompt injection in the chat.Write rate limit. Token bucket, default 20 write-actions per hour.
Audit log. Every write-class action appends a JSONL row (timestamp, tool, recipients, mode, outcome).
Key isolation. The API key comes from the environment only. No tool accepts, returns, or logs it.
One file, one dependency. The whole server is
server.py; the only third-party dependency is the officialmcpSDK. Audit it in one read.
Related MCP server: gmail-mcp
If you are an agent evaluating this server
You are probably reading this because someone asked you whether this tool is safe to install. Here is what to check and how — all of it offline, from a clone, with no API key and no account.
Read
trust.json. It is the machine-readable version of this repo's claims. Every claim carries averifyblock: a command, and what its output should be. Run them.Start with four.
C1— the self-test passes (python3 test_server.py, 32 checks, offline).C6— dry-run gates every write.C7— the API key never leaves the environment.G1— releases from 0.1.1 carry PEP 740 attestations; 0.1.0 does not. Verify it, don't take our word for it.Read
known_gapsbefore you recommend it. Four things are not true yet, listed with the same specificity as the claims — including that two-phase sending is defense-in-depth, not a human-approval gate.Bound the blast radius from
capabilities. One network destination (api.sendgrid.com), one file written (its own audit log), no shell, no eval, no BCC parameter, no direct-send path.
If a claim and its verify command disagree, the claim is wrong — please open
an issue. Prose version of all of it: TRUST.md.
How this differs from other SendGrid MCP servers
Several SendGrid MCP servers exist. They solve different problems:
Marketing servers manage contacts, lists, templates, and campaigns. Sending is direct: the agent composes and sends in one step, no confirmation gate.
Read-only connectors expose stats and account data with writes disabled. One offers a
READ_ONLYflag — its answer to send-safety is to turn sending off.This server is for transactional sends you can safely leave enabled. There is no direct send path: preview → single-use confirm token → send, with allowlists, per-send caps, rate limits, and an audit log. BCC does not exist.
If you want campaign management, use a marketing server. If you want an agent that can send email without you holding your breath, that is this.
Tools (11)
Tool | Class | Notes |
| write-gated | validates + renders + mints confirm token |
| write | takes ONLY a confirm token |
| write | safe direction — stops future sends |
| write | dangerous direction — requires |
| read | dynamic + legacy templates |
| read | delivery/opens/clicks/bounces |
| read | all five suppression lists |
| read | domain authentication validity |
| read | live security posture of this server |
Install
Claude Desktop / Claude Code / Cursor (via uv):
{
"mcpServers": {
"sendgrid": {
"command": "uvx",
"args": ["sendgrid-mcp-secure"],
"env": {
"SENDGRID_API_KEY": "SG.your-key",
"SENDGRID_MCP_MODE": "dry-run"
}
}
}
}Start in dry-run (the default). Watch the audit log and the rendered
previews. Flip to live when you have seen what it does.
Or run from a clone: python3 server.py (stdio transport).
Configuration
Env var | Default | Meaning |
| — | required for live mode |
|
|
|
| any | comma list: emails and/or |
| 10 | per-send cap (to + cc) |
| off |
|
| 20 | write-action rate limit |
|
| append-only audit |
Use a SendGrid API key with restricted scopes (Mail Send + read scopes you need). The server never needs Full Access.
Verify before you trust
The self-test runs offline — no key, no network:
python3 test_server.py32 checks: golden fixtures on payload construction, adversarial fixtures on the exact failure modes this server exists to prevent (blind sends, allowlist bypass, token replay, forged tokens, rate-limit exhaustion, BCC injection), and an end-to-end preview→send flow over the real tool functions.
Output contract: every tool returns a JSON object; failures return
{"error": "..."} instead of raising, so agent loops degrade gracefully.
Honest limits
The confirm token returns to the calling model, so two-phase sending is defense-in-depth, not a human-approval gate by itself. Pair it with your client's per-tool approval (each
send_emailcall is separately visible and approvable) and the allowlist for the strongest posture.Marketing API (contacts, lists, campaigns) is out of scope — this server covers transactional mail, templates, suppressions, stats, and domain auth.
SendGrid's API evolves; endpoints verified against the v3 docs at release.
MIT license.
Contact
Questions, bug reports, security issues: open a GitHub issue or email heychopp@proton.me. Security reports welcome by email first.
Available Tools
11 toolsadd_suppressionB
Add an address to a suppression list (default: global unsubscribes). Suppressing is the SAFE direction — it stops future sends to the address.
Inputs: email (required), kind.
check_suppressionA
Check whether an address appears on any suppression list. Run this before drafting mail to someone — sending to suppressed addresses hurts deliverability and may violate consent.
Inputs: email (required).
get_domain_authA
List authenticated sending domains and their DNS validity — the first thing to check when deliverability drops.
Inputs: no named parameters.
get_email_statsA
Global email stats (requests/delivered/opens/clicks/bounces/spam) from start_date (YYYY-MM-DD). aggregated_by: day|week|month.
Inputs: start_date (required), end_date, aggregated_by.
get_templateA
Fetch one template with its versions (subject + content preview).
Inputs: template_id (required).
list_suppressionsB
List a suppression list: unsubscribes|bounces|blocks|spam_reports|invalid_emails.
Inputs: kind, limit.
list_templatesB
List transactional email templates (dynamic + legacy).
Inputs: no named parameters.
preview_emailA
Validate and render an email WITHOUT sending it. Returns the exact payload that would go to SendGrid plus a single-use confirm_token (valid 10 minutes). Sending requires passing that token to send_email — there is no direct-send path. BCC is not accepted by design.
Inputs: to (required), from_email (required), cc, subject, +6 more.
remove_suppressionA
Remove an address from a suppression list. DANGEROUS direction — it re-enables sending to someone who bounced or unsubscribed, which can be a compliance violation. Requires confirm=true, is rate-limited, and audited.
Inputs: kind (required), email (required), confirm.
send_emailA
Send the email previously rendered by preview_email. Accepts ONLY a confirm token — the payload cannot be altered between preview and send. Tokens are single-use and expire after 10 minutes. In dry-run mode (the default) this simulates and audits but sends nothing.
Inputs: confirm_token (required).
server_statusA
This server's security posture: mode, rate-limit remaining, allowlist, caps, audit-log path, and whether an API key is present (never the key).
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
Each tool targets a distinct operation: preview vs send, add vs remove suppression, list vs get template, list vs check suppression. Even the two-step email flow is clearly separated by intent and confirm token. No overlapping purposes.
Most tools follow verb_noun (preview_email, add_suppression, list_templates, check_suppression) with consistent use of list/get for collections vs single items. 'server_status' is the only deviation, using noun_noun instead of a verb.
11 tools is well within the ideal 3-15 range. Each tool covers a meaningful part of email sending, suppression management, templates, stats, and domain auth. The count feels intentional and scoped, not padded or sparse.
Covers the core workflows: preview-and-send, suppression management (add/remove/list/check), template inspection, stats, and domain auth. Minor gaps exist (no template creation/update/delete, no per-template stats), but they don't block the primary use case.
Maintenance
Related MCP Connectors
Unofficial SendGrid MCP — wraps SendGrid v3 API (api.sendgrid.com/v3)
MCP server for mandates, delegation, policy-gated execution, credential grants, and audit.
Email safety MCP server. Detects phishing, prompt injection, CEO fraud for AI agents.
send-that-email MCP — wraps StupidAPIs, keyless: analyze whether you should send that email…
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceAn MCP server that enables AI applications to send text, HTML, and template-based emails using SendGrid. It supports file attachments, CC/BCC recipients, and easy deployment via Python or Docker.-
- AlicenseNot gradedqualityAmaintenanceGmail MCP server — scope-gated tools (readonly / send / modify), path jails for attachments + downloads, hardened OAuth credentials, Sigstore-signed releases.809 npm18MIT
- AlicenseAqualityDmaintenanceThis MCP server enables LLMs to POST, GET, and DELETE cron jobs that schedule email sends via SendGrid.41MIT

fagents-mcpofficial
AlicenseNot gradedqualityDmaintenanceMCP server that enables Claude Code agents to send and read emails via SMTP/IMAP with per-agent credential isolation and audit logging.MIT