[ Forwarding & ARC ]

Distribution groups, shared mailboxes and lists: trace the flow

Classify distribution, shared storage, replies and list redistribution before diagnosing the route, content and alignment behind an authentication result.

By the EmailAuth research team · 6 min read

[ Key points ]

  • Distribution and shared mailbox access are different operations.
  • A reply from a shared address is a new outgoing message, not automatically a forwarded copy.
  • Redistribution affects SPF and DKIM through different inputs.
  • A list-domain signature is useful for DMARC only when it aligns with the evaluated author domain.
On this page

Identify the operation behind the address

An address called a group can distribute copies, collect messages in a shared mailbox, or run a mailing-list workflow that edits and redistributes posts. Those operations have different authentication consequences. Classify the actual handling steps before asking why a group message failed DMARC. A familiar product label does not establish the outgoing envelope, signed content or receiver decision.

This guide compares proposed evidence, not inspected accounts. No group membership, shared mailbox, list configuration or original delivery sample was accessed. The fictional flows below illustrate how to ask better questions. Required original evidence: Direct and indirect delivery samples with route, envelope, signatures and trusted results compared.

Distinguish distribution from shared access

Microsoft describes Exchange Online distribution lists as a way to distribute messages and distinguishes them from dynamic groups, security groups and Microsoft 365 groups. Its shared-mailbox documentation instead describes several people accessing the same mailbox, with appropriate permissions for sending as or on behalf of that mailbox. These documented purposes do not predict how a particular tenant routes an external copy. Microsoft distribution lists Microsoft shared mailboxes

Separate receiving a message into shared storage from forwarding it elsewhere and from composing a reply. An employee opening a stored message has not thereby created another SMTP hop. A forwarding rule can create one. A reply sent using the shared address is a new outgoing message whose author domain and signing configuration need their own assessment. Record which operation the reported failure actually concerns.

Trace the envelope and message independently

At each SMTP boundary, record the connecting host and MAIL FROM identity separately from the visible RFC5322.From. SPF evaluates the sending connection against the relevant envelope identity. A redistributed message can retain the original visible author while reaching the next receiver from a different server. That difference is a reason to investigate SPF, not proof that every distribution mechanism must fail. RFC 7208 sections 2.4 and 4

DKIM follows a different dependency. A signature identifies its signing domain and covers selected header instances plus the canonicalized body, subject to any body-length limit. Merely sending the same content to additional recipients does not necessarily alter those inputs. Editing covered material may invalidate the signature, while a change removed by the selected canonicalization can leave verification unchanged. Inspect the actual h=, c= and any l= values. RFC 6376 sections 3.4–3.5

Recognize when a list becomes a mediator

A mailing-list manager may add a subject prefix, footer, list headers, digest packaging or a replacement author address. RFC 7960 documents these indirect-flow patterns and their interoperability problems. Use it as an explanation of possible transformations, not evidence that a named list enabled every feature. Capture the message before and after redistribution to establish which transformations occurred. RFC 7960 section 3.2.3

Evaluate every relevant DKIM signature independently. A list's additional signature does not repair an original signature whose covered content changed. For DMARC, a valid signature is useful only when its signing domain aligns with the author domain being evaluated. Likewise, SPF passing for a list-controlled envelope domain does not establish alignment with an unchanged external author domain. RFC 6376 section 4.2 RFC 9989 sections 3.2.10 and 5.3.5

Compare two fictional delivery branches

Imagine a fictional message from alex@writer.example addressed to discussion@club.example. One branch distributes an otherwise unchanged copy to an external member. Another list branch appends a footer after the original writer.example DKIM signature was added. Assume that signature originally verified, covers the full body and uses relaxed body canonicalization. These are explicit scenario assumptions, not observed results.

On the unchanged branch, the original signature can remain valid, subject to its other verification requirements. The added footer on the second branch changes the covered canonicalized body and invalidates that signature. If neither branch has aligned SPF, the first can still have an aligned DKIM path while the second needs a different explanation or mitigation. A fresh list-domain signature alone cannot align with the unchanged writer.example From address. RFC 6376 section 3.4.4 RFC 9989 section 5.3

Locate trustworthy authentication observations

Preserve the original raw copies and identify who inserted each Authentication-Results field. A result written at the organization's trusted ingress can describe the message before an internal modification. A downstream check describes the later representation. Both can be accurate at their respective boundaries. An arbitrary sender-supplied result header is not equivalent evidence; RFC 8601 requires a trust-boundary model for consuming these assertions. RFC 8601 section 7

If ARC appears, record its chain validation and the receiver's trust decision separately. RFC 8617 is Experimental and ARC is not a universal trust framework. The official proposal to reclassify it was an active Internet-Draft when reviewed on 8 November 2026, not an approved replacement specification. Neither an ARC header nor an accepted message proves that the original DKIM signature currently verifies. RFC 8617 section 9.3 IETF ARC status proposal

Choose a repair that fits the flow

The proposed decision should identify the actual failing branch: unauthorized envelope use, covered-content modification, incorrect author identity, or a receiver-specific trust decision. Then consider supported configuration changes with the service owner. Preserve required security processing and user expectations rather than disabling group features broadly. No permission to alter a group, tenant, signer or list is implied by this analysis.

An approved verification should compare a direct control with the relevant indirect branches, including external members and shared-address replies where in scope. Retain raw messages, timestamps, route evidence, signature details and results from known evaluators. Until those samples exist, the defensible output is a flow map and a test plan, not a declaration that the organization's groups are DMARC-safe.

[ Takeaway ]

An address label cannot tell you which authentication path survived. Map the actual operation and compare trusted observations on the relevant branches.

[ By role ]

What this means for you

Choose your role for practical next steps.

Messaging administrator

Classify operations before changing objects

A group object can participate in several routes, while shared storage and outbound composition have different dependencies.

Next steps

  1. Map receipt, redistribution and reply branches separately.
  2. Record which branches cross an external SMTP boundary.
  3. Verify actual configuration without guessing from the display label.

Authentication analyst

Follow each surviving identity

SPF authorization, DKIM integrity and DMARC alignment answer related but distinct questions.

Next steps

  1. Record each hop's envelope and connecting host.
  2. Evaluate individual signatures and their selected coverage.
  3. Associate results with a known evaluator and message representation.

Collaboration service owner

Preserve the intended member experience

A repair that changes posting or reply behavior needs more than an authentication pass.

Next steps

  1. Identify who should receive copies and who should send shared-address replies.
  2. Review proposed list modifications with the owner.
  3. Require controlled sample evidence before declaring the flow ready.

[ FAQs ]

Frequently asked questions

Does receiving mail in a shared mailbox create another SMTP hop?

Shared access to a stored message does not by itself create a forwarding transaction. A separate forwarding rule or redistribution step can do so. Determine whether the reported issue concerns receipt, onward forwarding or a newly composed shared-address reply.

Can I infer the outgoing envelope from the group name?

No. A group label describes neither the MAIL FROM used on an external hop nor the connecting server. Retain route and envelope observations for the actual branch rather than treating the object name as authentication evidence.

Does distributing an unchanged copy necessarily invalidate DKIM?

No. If the signature input remains equivalent under its canonicalization and other verification requirements hold, redistribution can preserve it. Inspect signed header instances, body coverage and actual changes instead of assuming every extra recipient causes failure.

Is a list footer always harmless under relaxed canonicalization?

No. Relaxed body canonicalization handles specified whitespace changes, not arbitrary added wording. A footer that changes covered canonicalized body content invalidates that previously valid signature. Coverage and the actual transformation still need inspection.

Does the list's own DKIM pass authenticate an unchanged external From domain?

Only if its signing domain aligns with the author domain under the applicable DMARC rules. A valid signature from an unrelated list domain does not supply that alignment, and it does not restore a broken original signature.

Why can an ingress result differ from the final recipient's check?

They may describe different representations at different stages. Preserve the evaluator, trust boundary and message version for each result. A sender-supplied Authentication-Results field cannot simply be trusted as the ingress assessment.

Does ARC guarantee that every receiver accepts list mail?

No. ARC validation does not establish every handler's trustworthiness or guarantee disposition. RFC 8617 is Experimental; the reviewed Historic reclassification proposal is still an active draft, not a completed standards change.

What would establish the behavior of our actual group?

Collect approved direct and indirect delivery samples with route, envelope, signatures and trusted results compared. Include the branches in question, such as external members or shared-address replies. No tenant or delivery evidence was collected here.

[ Sources ]

Sources

  1. IETF Datatracker: Reclassifying ARC as Historic
  2. Microsoft Exchange Online: Distribution lists
  3. Microsoft 365: About shared mailboxes
  4. RFC 6376: DomainKeys Identified Mail (DKIM) Signatures
  5. RFC 7208: Sender Policy Framework (SPF) for Authorizing Use of Domains in Email, Version 1
  6. RFC 7960: Interoperability Issues between DMARC and Indirect Email Flows
  7. RFC 8601: Message Header Field for Indicating Message Authentication Status
  8. RFC 8617: The Authenticated Received Chain (ARC) Protocol
  9. RFC 9989: Domain-based Message Authentication, Reporting, and Conformance (DMARC)

[ Glossary ]

Glossary

Distribution branch
A path by which a message addressed to a group is delivered onward to a member; its actual routing and modifications require evidence.
Shared mailbox access
Authorized access by multiple people to the same mailbox, distinct from a separate onward SMTP delivery.
Mailing-list mediator
A handler that accepts and redistributes list posts, potentially applying list-specific transformations.
Author domain
The domain extracted from RFC5322.From for the DMARC evaluation of the message being assessed.
Trusted evaluator
An authentication-results producer whose assertions the receiving system accepts under an established trust-boundary policy.

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