Email authentication rollout for MSPs

A client can have valid DNS records and still send unauthenticated mail. Use this checklist to connect domain ownership, authoritative DNS, sending services and message evidence before you tighten DMARC.

1. Inventory owners and senders

The registrar sells the domain. The authoritative DNS host publishes its records. The mail provider signs messages. These may be three different services. Check delegated nameservers before choosing where to write, and obtain the client's authorization.

List every legitimate source: employee mail, CRM, billing, support, newsletters, website forms and other transactional senders. Ask each owner which From domain, return-path domain and DKIM signing domain the service uses. Do not remove an SPF include just because its name is unfamiliar.

Reusable rollout worksheet

This is an illustrative worksheet for example.com, not a customer result. Keep real client details in your authorized internal system, not a public prompt or URL. Create one sender row per service; a single domain can have several.

FieldOwner / actionEvidence to retain
Domain and authorizationClient owner approves work on example.comAuthorized scope and rollback contact
Authoritative DNS hostDNS administrator confirms nameservers and accessCurrent full record set, TTLs and provider connection
Sending serviceMail owner records the service and From domainRepresentative message headers and current provider instructions
SPF authorizationSender owner supplies required mechanismsRecursive evaluation, including nested includes and redirects
DKIM selector / public recordMail owner supplies DNS record and enables signingPublished selector plus a verified signed message; never the private key
From-domain alignmentMail owner checks authenticated identifiersAligned SPF or DKIM pass for each legitimate source
Reporting destinationSecurity owner configures and reviews reportsReceived aggregate reports and unexplained failure sources
Planned change / verificationOperator records proposed diff and policy choiceApproval, observed DNS propagation and post-change messages
RecoveryNamed rollback owner retains prior valuesRecovery procedure and expected DNS/policy cache delays

2. Verify SPF, DKIM and alignment

SPF: count evaluation work, not vendors

SPF authorizes sending hosts for the envelope sender domain. RFC 7208 limits lookup-causing terms during evaluation to ten, including nested include and redirect evaluation. A short top-level record can still exceed the limit. IP address mechanisms and all do not consume that term limit.

Keep one applicable SPF record and preserve legitimate senders. Remove obsolete authorization only after confirming ownership and use. Flattening is not a permanent one-click fix: provider addresses can change, and incomplete conversion can lose mechanisms.

DKIM: publish the public key, then verify signing

Your mail provider generates or manages the signing key and supplies a selector and public DNS record. Publish it at the specified name, enable signing in the mail system, and inspect a real message. A common-selector DNS probe is not exhaustive and does not prove signing, key strength or alignment. Never send a DKIM private key to Authwright or an LLM.

DMARC: connect authentication to the visible From domain

DMARC requires an authenticated SPF or DKIM identifier to align with the author's From domain. Passing SPF for an unrelated return-path domain alone is not enough. A p=none policy requests monitoring; it is not a broken record, but it does not request quarantine or rejection.

Gmail and Yahoo have sender-category requirements beyond DNS authentication. Gmail's higher-volume requirements apply around 5,000 messages per day to personal Gmail accounts; Yahoo defines its own bulk category. Read their current guidance rather than treating any DNS grade as full compliance.

3. Stage changes by evidence, not a fixed calendar

  1. Capture current records, sender owners and a recovery contact. Verify access to the actual DNS host.
  2. Publish missing legitimate sender records and enable DKIM signing at each sender. Inspect messages from each service.
  3. Configure a suitable reporting destination and review received aggregate reports. Missing reports are not evidence of no traffic.
  4. Resolve unexplained failures, forwarding effects and misaligned third-party senders. Agree the intended policy with the client.
  5. Move a verified domain or domain group toward quarantine or reject when evidence supports it. Observe after each change and retain a recovery path.

There is no universal “wait 24 hours, then reject” schedule. Current DMARC core RFC 9989 removes the older pct sampling tag. Stage by verified domains and sender readiness rather than prescribing percentage ramps. DNS caching means restoring a record does not make every receiver see it immediately.

4. Use Authwright deliberately

Start with the free published-DNS check. Its SPF count is top-level only, its DKIM probe covers common selectors, and its MTA-STS check finds TXT discovery only. Use mail-provider evidence for the missing pieces.

For MCP work, call email_auth_diagnose, then email_auth_propose with explicit sender and policy choices. Review every operation before email_auth_apply, then inspect wait_for_propagation results. The combined email_auth_wizard is write-capable; it is not a promise of a separate confirmation screen.

Use skipBimi: true for proposals until real logo and certificate prerequisites are prepared: end-to-end BIMI validation is not implemented. DKIM signing remains the mail provider's job. Email-auth apply, wizard, hosting, flattening and report access require Pro or above; bulk tools require Team or above.

Check provider-specific details for GoDaddy, Namecheap replace-all DNS, Cloudflare mail records, Porkbun and Route 53. The capabilities page explains operation boundaries. Receiving-domain transport policy has its own MTA-STS setup checklist.

Troubleshooting

Primary references