@dmarc-examiner/mcp
MCP serverFiles & storageYour AI can read the DMARC reports for your own domain, covering both aggregate and forensic reports. That means you can ask it to show which sources have been sending email as your domain, whether their messages align with your settings, and where alerts show up in the reports. Findings can be exported as CSV.
Available today. Use it from your connected AI after setup.
No other account needed.
Add the app and ask your AI to read the DMARC reports for your domain. From there, request a summary, drill into a specific source, or get a CSV export of the findings.
Then ask your AI: use @dmarc-examiner/mcp to list domains
What your AI can do with it
- Read your DMARC aggregate reports
- Go through forensic DMARC reports
- Show which sources are sending email as your domain
- Check whether messages align with your DMARC settings
- Spot alerts flagged in your reports
- Export report data as CSV
From the project's README
As published by dmarc-examiner/mcp in README.md.
Official MCP (Model Context Protocol) server for DMARC Examiner. Query your DMARC monitoring data from AI assistants like Claude Desktop, Claude Code, and Cursor.
Why
DMARC aggregate reports arrive as compressed XML, one file per receiver per day. Reading them means either opening a dashboard or parsing XML by hand.
This server puts that data behind an MCP connection, so you can ask questions instead:
Which sending sources failed DMARC alignment last week?
Show me the domains where SPF passes but DKIM does not.
Export the report for example.com as CSV.
DMARC Examiner is a DMARC monitoring service with a free tier — you need an account to use this server, but not a paid one.
Quick Setup
Option 1: Remote URL (Recommended)
Most MCP clients support remote HTTP servers directly. Add this to your MCP configuration:
{
"mcpServers": {
"dmarc-examiner": {
"url": "https://mcp.dmarc-examiner.com/mcp"
}
}
}
Option 2: npx (for clients without remote HTTP support)
{
"mcpServers": {
"dmarc-examiner": {
"command": "npx",
"args": ["-y", "@dmarc-examiner/mcp"]
}
}
}
Configuration by Client
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"dmarc-examiner": {
"url": "https://mcp.dmarc-examiner.com/mcp"
}
}
}
Claude Code
claude mcp add dmarc-examiner --transport http https://mcp.dmarc-examiner.com/mcp
Cursor
Add to your Cursor MCP settings:
{
"mcpServers": {
"dmarc-examiner": {
"url": "https://mcp.dmarc-examiner.com/mcp"
}
}
}
Authorization
Adding the server does not require authorizing. It starts immediately and answers initialize, tools/list and ping without credentials, so your client can list all 21 tools straight away.
A browser opens the first time you call a tool that touches your data. You'll need to:
- Log in to your DMARC Examiner account
- Select which organization to connect
- Approve the requested permissions (scopes)
The call you made is then replayed automatically, so it completes rather than failing.
With Option 1 the MCP client runs this flow itself. With Option 2 the package does it: it registers as an OAuth client, opens your browser, and stores the resulting tokens in ~/.dmarc-examiner/mcp-credentials.json with 0600 permissions. The refresh token is reused on later runs, so the browser step happens once.
Why it works this way. Authorizing on startup is the obvious design and it breaks every automated client: directory checks, scanners and CI have no browser to open, so the server appears to hang and the check times out. We wrote up the failure and the fix, with logs, in Your Remote MCP Server Will Fail Every Directory Check.
You can also drive it directly:
npx -y @dmarc-examiner/mcp login # authorize now, outside your MCP client
npx -y @dmarc-examiner/mcp logout # forget the stored credentials
Set DMARC_EXAMINER_MCP_URL to point the package at a different endpoint.
Available Tools
Domains
| Tool | Description | Scope |
|---|---|---|
list_domains | List all monitored domains in your organization | domains:read |
get_domain | Get details of a specific domain | domains:read |
check_dmarc | Inspect the public DMARC record at _dmarc.<domain>. Works on any domain, not only your own | domains:read |
create_domain | Add a domain. Returns the reporting email and DMARC record to publish in DNS | domains:manage |
verify_domain | Re-check DNS to confirm the DMARC record carries your reporting address | domains:manage |
delete_domain | Remove a domain from monitoring. Soft delete — historical reports remain | domains:manage |
Aggregate reports (RUA)
| Tool | Description | Scope |
|---|---|---|
list_reports | List DMARC reports with filtering | reports:read |
get_report | Get a detailed report with records | reports:read |
get_report_statistics | Report statistics by country and ASN | reports:read |
export_report_csv | Export a report as CSV | reports:export |
Forensic reports (RUF)
Requires the Pro plan or above; these return 403 otherwise.
| Tool | Description | Scope |
|---|---|---|
list_forensic_reports | List RUF reports with filtering and pagination | forensic:read |
get_forensic_report | Full detail of a forensic report, including the failed message envelope | forensic:read |
get_forensic_reports_statistics | Counters by failure type, delivery result, source IP with geolocation, and domain | forensic:read |
get_forensic_reports_summary | Forensic activity over the last 7 days | forensic:read |
Alerts
| Tool | Description | Scope |
|---|---|---|
list_alerts | List security alerts | alerts:read |
dismiss_alert | Dismiss an alert | alerts:manage |
Webhooks
Requires the Pro plan or above; these return 403 otherwise.
| Tool | Description | Scope |
|---|---|---|
list_webhooks | List configured webhook endpoints | webhooks:read |
create_webhook | Create an endpoint. The URL must be HTTPS | webhooks:manage |
update_webhook | Update the URL, name or active state of an endpoint | webhooks:manage |
delete_webhook | Delete an endpoint permanently | webhooks:manage |
test_webhook | Send a webhook.test event and return the delivery status | webhooks:manage |
Scopes
| Scope | Description |
|---|---|
domains:read | View monitored domains |
domains:manage | Add, verify and remove domains |
reports:read | View reports and statistics |
reports:export | Export reports to CSV |
forensic:read | View forensic (RUF) reports |
alerts:read | View alerts |
alerts:manage | Manage alerts (dismiss) |
webhooks:read | View webhook endpoints |
webhooks:manage | Create, update, delete and test webhooks |
Documentation
For detailed documentation, visit: https://dmarc-examiner.com/docs/settings/mcp-integration?utm_source=github&utm_medium=content&utm_campaign=mcp-repo
Support
- Documentation: https://dmarc-examiner.com/docs?utm_source=github&utm_medium=content&utm_campaign=mcp-repo
- Email: support@dmarc-examiner.com
License
MIT — see LICENSE.
Tools it offers (21)
What this server listed when ahel dialed its public endpoint in Sep 2026, with no key and no account of yours. The names are the server’s own.
list_domainsget_domaincheck_dmarccreate_domainverify_domaindelete_domainlist_reportsget_reportget_report_statisticsexport_report_csvlist_alertsdismiss_alertlist_forensic_reportsget_forensic_reportget_forensic_reports_statisticsget_forensic_reports_summarylist_webhookscreate_webhookupdate_webhookdelete_webhooktest_webhook
Signals
- Last commit
- Aug 2026
Advanced
- Delivery
- mcp MCP server → your ahel gateway (mcp.ahel.ai) → every connected AI client.
- Catalog kind
- mcp-server
- Gateway key
io-github-dmarc-examiner-mcp- Source
- github.com/dmarc-examiner/mcp
- Hosted endpoint
https://mcp.dmarc-examiner.com/mcp