[ Reports & investigation ]

Separate forwarding, direct sending and suspected spoofing

A route-based investigation guide that keeps authentication results, forwarding effects, ARC trust and receiver disposition distinct.

By the EmailAuth research team · 7 min read

[ Key points ]

  • Forwarding changes the connecting address and may change the envelope identity.
  • An unchanged aligned DKIM signature can survive forwarding, while modifications can break it.
  • ARC validation does not establish sealer trust or message safety.
  • No actual route map or controlled message has been retained for the fictional alternatives.
On this page

Classify the route before deciding what failed

A receiver can see mail directly from an originating service, after a forwarding hop or after a mailing list has modified it. A source IP and authentication failure do not, by themselves, distinguish these paths from impersonation. The investigator must connect the reported result to a particular route and preserve uncertainty where the route is missing. Otherwise a repair intended for a legitimate sender can authorize unrelated infrastructure.

This guide uses RFC 9989 and RFC 9990 from May 2026, SPF RFC 7208 and the experimental ARC specification RFC 8617. Provider guidance was reviewed on 8 November 2026. All route examples are fictional paper cases. No message was sent, imported or verified, no DNS query was run and no forwarding or receiver-trust setting was changed.

Define direct sending relative to the receiving boundary

For this investigation, direct sending means that the intended outbound service delivers to the receiving boundary without an intervening forwarding service in the route under examination. It does not mean that the originating organization uses only one internal server. Internal relays and outbound gateways can precede the observed connection, so record the boundary used when applying the label.

RFC 7208 evaluates a connecting host against a defined SMTP identity. In a direct route, a failure can indicate a configuration problem, an unexpected path or unauthorized use. It is not sufficient to name the cause from the result alone. Preserve the evaluated domain, connection context and error status before proposing any change to its authorization.

Explain why forwarding can alter SPF evidence

When a forwarder delivers an original message onward, the next receiver sees the forwarder's connecting address. If the original envelope identity is retained, that address may not be authorized by the original domain's SPF policy. If the envelope is rewritten, SPF may instead authenticate the forwarder's identity. Neither circumstance automatically provides an aligned SPF path for the original visible From domain under RFC 9989.

That is why adding every observed forwarder to the original domain's SPF is not a sound default response. It can widen authorization without addressing the actual route or the other authentication path. First establish whether forwarding occurred, whether the recipient expected it and which identities changed. A forwarding-like pattern is a hypothesis until representative route evidence supports it.

Examine whether the original DKIM signature survived

A message can retain a valid aligned DKIM signature through forwarding if the signed material remains compatible with verification. RFC 6376 defines what the signature covers and how it is checked. Changes to signed headers or body content can break that signature; merely moving the message through another server does not necessarily do so. Do not generalize from one forwarding product to all routes.

A mailing list can alter content, add a footer or otherwise transform the message before redistribution. A new valid signature from the list may authenticate the list's domain while remaining unaligned with the original From domain. A reported DKIM pass therefore needs its signing identity and alignment context. It cannot be read as proof that the original author's domain passed DMARC.

Use ARC as evidence whose trust still needs a decision

RFC 8617 is an experimental protocol for carrying authentication assessments across message-handling steps. ARC chain validation and trust in the sealing parties are separate questions. A cryptographically valid chain does not make the message safe or establish that every sealer is trustworthy. A missing or broken chain also does not prove that the message was forged.

Receivers can use ARC information in local disposition decisions. Do not rewrite that into a universal rule that ARC turns an underlying SPF or DKIM alignment failure into a protocol pass. Preserve the observed authentication results and any reported policy override separately. RFC 9990 distinguishes DMARC-evaluated fields from raw authentication results and provides context for policy-override reporting. Receiver acceptance and authentication outcome remain different facts.

Compare four fictional paths without overclassifying

The retained paper exercise contrasts four possibilities: an intended direct service; an unchanged forwarder preserving aligned DKIM; a modifying list whose original signature no longer verifies; and a sender impersonating the From domain without a known authorized route. These are explanatory alternatives, not four tested messages. No real host, account or business owner has been assigned to them.

The first three possibilities require route-specific configuration and message evidence before they can be treated as legitimate observed traffic. The fourth remains suspected spoofing until the available evidence and security review support that classification. Authentication failure alone cannot distinguish the modifying-list case from impersonation. Conversely, passing authentication cannot establish benign content or a currently approved business purpose.

Respect the receiver's actual trust configuration

Microsoft's trusted ARC sealer guidance is a provider-specific example: it describes selecting legitimate, required intermediaries for Microsoft 365 receiving behavior. It does not establish a universal allowlist, another provider's policy or the configuration of a customer tenant. This draft neither adds a sealer nor recommends trusting an intermediary merely because its name is familiar.

For a future investigation, retain the representative message, trusted receiving trace, relevant service configuration and recipient's expected route. Authentication-Results fields must be interpreted within an established receiving trust boundary, as RFC 8601 explains. An untrusted copied header can assert anything. Keep message disclosure, receiver-security changes and production DNS actions subject to their own human approvals.

Require an observed route before calling the remedy proven

The exact intake gate remains open: Validate current primary specifications and retain a worked example before drafting. Primary sources and fictional comparisons have been retained, but the commissioned original evidence is still absent: A route-to-requirement map with current provider evidence and a controlled test message.

An eventual controlled test must use an approved recipient and route, retain the relevant diagnostics and distinguish delivery, authentication and local-policy handling. It should address the specific uncertainty rather than promise universal forwarding compatibility. Until that evidence exists, preserve the candidate classification and next investigation step. Do not claim successful remediation, confirmed spoofing or an approved sender from the documentary examples alone.

[ Takeaway ]

Explain the observed path before calling a failure spoofing or proposing broader sender authorization.

[ By role ]

What this means for you

Choose your role for practical next steps.

Mail-flow investigator

Establish the path before the verdict

Keep candidate direct, forwarded, modified and unauthorized paths distinct until the message and receiver evidence support classification.

Next steps

  1. Record the receiving boundary and identity changes.
  2. Preserve trusted trace and per-hop evidence where available.
  3. Do not turn a failing aggregate row into a confirmed spoofing allegation.

Receiving-system administrator

Separate ARC validation from trust

Own any receiver-specific acceptance policy and verify the actual intermediary rather than accepting a valid chain as sufficient.

Next steps

  1. Review the legitimate service and expected modifications.
  2. Retain underlying authentication outcomes alongside local-policy handling.
  3. Require separate approval before changing trusted-sealer or other security settings.

Sending-service owner

Authorize a scoped validation

Approve the route and harmless test needed to resolve the specific uncertainty without broadening sender permissions by default.

Next steps

  1. Confirm whether the indirect path is expected by the recipient.
  2. Keep external disclosure and DNS changes separately authorized.
  3. Accept only the observed result and its limitations, not universal forwarding compatibility.

[ FAQs ]

Frequently asked questions

Does direct sending mean there are no internal relays?

No. Here it is defined relative to the receiving boundary and the forwarding route under investigation. Internal or outbound components can precede the observed connection without making every hop an external forwarder.

Why can unchanged envelope forwarding cause an SPF failure?

The next receiver sees the forwarder’s connecting IP, which may not be authorized for the retained original SMTP identity. Preserve the identity and route before attributing the failure to forgery.

Does rewriting the envelope automatically fix DMARC alignment?

No. SPF can authenticate the rewritten identity while that domain remains unaligned with the original visible From domain. The other aligned authentication path must be considered separately.

Must every forwarder break the original DKIM signature?

No. Verification can survive when the signed material remains compatible with the signature. Changes to signed headers or body content can invalidate it; the actual message and route determine the result.

Does arc=pass mean the message or sealer is trustworthy?

No. Chain validation and trust are different questions. RFC 8617 does not make a safety assessment, and even a validated chain can accompany malicious content or an untrusted handling actor.

Does missing ARC establish that the message was spoofed?

No. A route may lack ARC participation, and validation can fail for reasons other than forgery. Keep that evidence separate from the security classification and investigate the path.

Should every familiar forwarder become a trusted Microsoft 365 sealer?

No. Microsoft’s guidance limits that decision to legitimate, required services. It is receiver-specific configuration, not permission to trust a brand or apply the same policy to other providers.

What remains necessary before a forwarding remedy is called verified?

An approved route-to-requirement map and controlled message, with provider evidence, relevant diagnostics and a clear outcome. Neither the fictional comparisons nor a documentary ARC explanation supplies that test.

[ Sources ]

Sources

  1. Microsoft: Configure trusted ARC sealers
  2. RFC 6376: DomainKeys Identified Mail Signatures
  3. RFC 7208: Sender Policy Framework
  4. RFC 8601: Authentication-Results Header Field
  5. RFC 8617: Authenticated Received Chain Protocol
  6. RFC 9989: Domain-Based Message Authentication, Reporting, and Conformance
  7. RFC 9990: DMARC Aggregate Reporting

[ Glossary ]

Glossary

Receiving-boundary route
The defined message path being investigated up to a particular receiver, making clear which relays and forwarding steps are in scope.
Envelope rewrite
A change to the SMTP return-path identity during handling, which may alter the SPF-authenticated domain without aligning it to the original visible From.
ARC chain validation
The protocol assessment of the carried ARC sets, distinct from a receiver’s trust in the sealers and from message safety.
Local disposition decision
The receiver’s handling choice, which can use additional evidence and must not be confused with rewriting an underlying authentication result.
Suspected spoofing
A provisional security classification for evidence suggesting unauthorized identity use, not an automatic consequence of a missing owner or failed authentication check.

Enterprise provenance

Why top financial institutions and governments retired their email security dashboards.

Traditional email authentication platforms sell you a dashboard and ask your engineers to do the labour. In high-stakes environments, manual oversight means delayed enforcement, configuration drift and unmonitored shadow domains.

EmailAuth is the agentic email provenance platform. Our agents enforce p=reject, keep third-party senders aligned and move your domains to full RFC 9989 compliance, the new DMARC standard,, with human-in-the-loop auditability.

Governments Large enterprises Tier-1 financial institutions

SaaS · Dedicated tenancy · Air-gapped on-prem · BYOK encryption

Knowledge Base