]> Stalwart Labs LLC

1309 Coffeen Avenue, Suite 1200 Sheridan WY 82801 USA mauro@stalw.art https://stalw.art

Internet-Draft This document specifies MTA Hooks, an HTTP-based protocol enabling Mail Transfer Agents (MTAs) to delegate message processing decisions to external services. MTA Hooks provides a modern alternative to legacy mail filtering protocols by leveraging ubiquitous HTTP infrastructure, supporting both JSON and CBOR serialization, and offering fine-grained capability negotiation. The protocol supports both inbound message reception and outbound message delivery scenarios, allowing external scanners to inspect messages, modify content, and influence routing decisions at various stages of mail processing.

Introduction Mail Transfer Agents require extensible mechanisms to integrate with external services for spam filtering, virus scanning, policy enforcement, and compliance monitoring. While legacy protocols such as the Milter protocol have served this purpose, they present challenges in modern deployments including proprietary wire formats, limited tooling, and operational complexity. MTA Hooks addresses these challenges by defining an HTTP-based protocol for mail processing delegation. The protocol leverages the widespread availability of HTTP client and server implementations, standard serialization formats, and established operational practices for HTTP services. MTA Hooks achieves broad deployment compatibility by building upon standard HTTP semantics and methods, allowing implementers to leverage existing HTTP client and server libraries. The protocol supports both JSON and CBOR serialization to accommodate environments with different performance and parsing requirements. Capability negotiation during registration enables graceful feature discovery and forward compatibility as the protocol evolves. Mandatory transport encryption protects message content in transit, while the authentication mechanism remains pluggable to integrate with diverse deployment environments. The HTTP foundation ensures compatibility with existing operational infrastructure including load balancers, reverse proxies, and monitoring systems. MTA Hooks supports two primary use cases: inbound processing during message reception from remote clients, and outbound processing during message delivery to remote servers. Both scenarios follow a consistent request-response pattern where the MTA invokes the registered hook endpoints at configured processing stages.

Notational Conventions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.

Terminology The following terms are used throughout this document: MTA (Mail Transfer Agent): A software component responsible for transferring electronic mail messages between hosts, as described in . Hook Endpoint: An HTTP server endpoint that receives hook invocations from an MTA and returns processing instructions. Scanner: A service that registers with an MTA to receive hook invocations. Scanners perform functions such as spam filtering, virus scanning, or policy enforcement. The terms "scanner" and "hook endpoint" are used interchangeably when referring to the service receiving hook requests. Registration: The process by which a scanner establishes a relationship with an MTA, including capability negotiation and callback URL configuration. Inbound Processing: Hook invocation during message reception, when the MTA accepts a message from a remote SMTP client. Outbound Processing: Hook invocation during message delivery, when the MTA transmits a message to a remote SMTP server. Stage: A specific point in mail processing at which the MTA invokes registered hooks. Different stages provide access to different message properties. Action: An instruction from a scanner indicating how the MTA should proceed with message processing. Modification: A change to message properties requested by a scanner, expressed as operations on the hook request object.

Data Types This specification defines several object types used throughout the protocol. The following primitive types are used:

String

A JSON string value.

Int

An integer in the range -2^53+1 <= value <= 2^53-1.

UnsignedInt

An integer in the range 0 <= value <= 2^53-1.

Boolean

A JSON boolean value (true or false).

UTCDate

A string in "date-time" format as defined in , where the time-offset component MUST be "Z" (UTC time). For example, "2024-12-21T15:30:00Z".

Protocol Overview MTA Hooks defines a request-response protocol where the MTA acts as an HTTP client and scanners act as HTTP servers. During mail processing, the MTA constructs a request containing message properties and context information, transmits it to registered scanner endpoints, and processes the response to determine subsequent actions.

Architecture The protocol architecture consists of three primary components: The MTA, which initiates HTTP requests to scanner endpoints at configured processing stages. Scanner services, which receive requests, perform analysis, and return responses containing actions and modifications. A registration mechanism through which scanners declare their capabilities and subscribe to specific processing stages.

| | | | 2. Discovery Response | | | |<-----------------------------| | | | | | | | 3. Registration Request | | | MTA |----------------------------->| Scanner | | | 4. Registration Response | | | |<-----------------------------| | | | | | | | 5. Hook Request | | | |----------------------------->| | | | 6. Hook Response | | | |<-----------------------------| | +-------+ +---------+ ]]>

The protocol flow proceeds as follows: The MTA discovers scanner capabilities by requesting the well-known discovery endpoint. The scanner returns a discovery document describing supported stages, actions, and configuration options. The scanner submits a registration request specifying desired stages, properties, and callback configuration. The MTA validates the request, optionally verifies the callback endpoint, and returns registration confirmation. During mail processing, the MTA invokes the scanner's callback URL with message properties and context. The scanner returns a response containing an action and optional modifications.

Serialization Requests and responses are serialized using either JSON or CBOR . The serialization format is negotiated during registration and indicated via HTTP Content-Type headers. For JSON serialization, the Content-Type header MUST be "application/json". For CBOR serialization, the Content-Type header MUST be "application/cbor". Binary data handling differs between formats. In JSON serialization, binary content such as message body parts MUST be encoded using Base64. In CBOR serialization, binary content SHOULD be transmitted as raw byte strings.

Processing Stages MTA Hooks defines processing stages for both inbound and outbound message handling. Scanners subscribe to specific stages during registration and receive invocations only for subscribed stages.

Inbound Stages Inbound stages correspond to SMTP protocol events during message reception: connect: Invoked when a remote client establishes a connection, before any SMTP commands are exchanged. ehlo: Invoked after the client sends EHLO or HELO command. mail: Invoked after each MAIL FROM command. rcpt: Invoked after each RCPT TO command. This stage may be invoked multiple times per transaction. data: Invoked after the complete message has been received, following the DATA command and message content.

Outbound Stages Outbound stages correspond to delivery events during message transmission: delivery: Invoked after a delivery attempt completes for all recipients, regardless of success or failure. defer: Invoked only when at least one recipient delivery cannot be completed and requires retry. dsn: Invoked when the MTA generates a Delivery Status Notification of any type. For outbound stages, the hook is invoked once after delivery attempts for all recipients in a given delivery batch have completed. The scanner receives the delivery status for each recipient and may modify per-recipient handling.

Inbound Message Processing During inbound processing, the MTA invokes registered hooks as it receives messages from remote SMTP clients. This corresponds to traditional mail filtering scenarios where external services inspect incoming mail for spam, viruses, or policy violations. The inbound flow proceeds through SMTP protocol stages: A remote client connects to the MTA. The MTA invokes hooks registered for the "connect" stage. The client sends EHLO/HELO; the MTA invokes "ehlo" stage hooks. The client sends MAIL FROM; the MTA invokes "mail" stage hooks. The client sends one or more RCPT TO commands; the MTA invokes "rcpt" stage hooks for each. The client sends DATA and message content; the MTA invokes "data" stage hooks. At each stage, scanners may instruct the MTA to accept, reject, or modify the transaction.

Multiple Scanner Handling When multiple scanners are registered for the same stage, the MTA invokes them sequentially in an implementation-defined order. Each scanner's response affects the request seen by subsequent scanners. The scanner chain terminates early when a scanner returns a terminal action: Inbound terminal actions: "reject", "discard", "disconnect" Outbound terminal actions: "cancel" Non-terminal actions ("accept", "quarantine" for inbound; "continue" for outbound) allow the chain to proceed, and subsequent scanners may override these actions. Implementations SHOULD provide configuration options for: Scanner invocation order (priority-based or explicit ordering) Whether to allow subsequent scanners to override "quarantine" decisions Logging of chain termination events for operational visibility Example chain behavior: Scanner A (spam filter) sets action to "accept" and adds X-Spam-Score header Scanner B (virus scanner) receives modified request, detects malware, sets action to "reject" Chain terminates; Scanner C is not invoked MTA rejects the message

Outbound Message Processing During outbound processing, the MTA invokes registered hooks as it delivers messages to remote SMTP servers. This supports use cases including delivery logging, compliance monitoring, and routing decisions. The outbound flow proceeds as follows: The MTA selects a queued message for delivery and identifies the recipients due for delivery at this time. The MTA attempts delivery to all selected recipients. After all attempts in this delivery job complete, the MTA invokes hooks registered for the "delivery" stage. If any recipient requires retry, the MTA invokes "defer" stage hooks. If the MTA generates a DSN, it invokes "dsn" stage hooks before sending. A delivery job represents a single processing cycle where the MTA retrieves a message from the queue and attempts delivery to one or more recipients. The specific batching of recipients into delivery jobs is implementation-defined and may depend on factors such as destination domain, connection reuse, or queue configuration. The "delivery" stage hook is invoked once per delivery job after all recipient attempts within that job have completed. Scanners may modify per-recipient delivery status, suppress DSN generation, or alter retry scheduling.

Modifications Scanners communicate changes through modifications to the hook request object. Modifications are expressed as operations on JSON Pointers referencing properties within the request structure. Three modification operations are supported: Set: Replace the value at a specified path. Add: Insert a value at a specified path, including array element insertion. Delete: Remove the value at a specified path. The MTA applies modifications in a defined order: set operations first, followed by add operations, and finally delete operations. Any modification that fails (for example, referencing a non-existent path for deletion) SHOULD be logged and included in failure statistics, but MUST NOT cause the entire response to be rejected. If a scanner modifies both the "message" and "rawMessage" properties, the "rawMessage" modification takes precedence and "message" modifications are ignored.

Discovery MTAs discover scanner capabilities through a well-known HTTP endpoint. Discovery is OPTIONAL; scanner endpoints MAY alternatively be configured manually.

Well-Known Endpoint Scanners that support discovery MUST expose a discovery document at the path "/.well-known/mta-hooks". The MTA retrieves this document using an HTTP GET request. The discovery endpoint MUST support JSON serialization. Support for CBOR serialization is OPTIONAL.

Discovery Document Format The discovery document is a JSON or CBOR object containing the following fields:

version: String

The protocol version. This specification defines version "1.0".

endpoints: DiscoveryEndpoints

URLs for protocol operations.

serialization: String[]

Supported serialization formats. Valid values are "json" and "cbor".

capabilities: DiscoveryCapabilities

Supported protocol capabilities.

limits: DiscoveryLimits|null

Operational limits, or null if no specific limits are advertised.

extensions: String[]|null

Supported protocol extensions. Extension identifiers SHOULD use a reverse domain name prefix for vendor-specific extensions.

A DiscoveryEndpoints object has the following properties:

registration: String

URL path for submitting registration requests.

deregistration: String

URL path template for deregistration, with "{registration_id}" placeholder.

A DiscoveryCapabilities object has the following properties:

inbound: DirectionCapabilities|null

Inbound processing capabilities, or null if inbound processing is not supported.

outbound: DirectionCapabilities|null

Outbound processing capabilities, or null if outbound processing is not supported.

A DirectionCapabilities object has the following properties:

stages: String[]

Supported stages for this processing direction.

actions: String[]

Supported actions that scanners may request.

fetchProperties: String[]

JSON Pointers for properties the MTA can include in hook requests.

updateProperties: String[]

JSON Pointers for properties scanners may modify.

A DiscoveryLimits object has the following properties:

maxMessageSize: UnsignedInt|null

Maximum message size in bytes the scanner accepts.

maxRegistrations: UnsignedInt|null

Maximum concurrent registrations.

timeoutMs: UnsignedInt|null

Default timeout in milliseconds for hook invocations.

The fetchProperties and updateProperties arrays contain JSON Pointers as defined in . Each pointer identifies a property within the hook request structure that the MTA supports including or that scanners may modify. For example, "/envelope/from/address" refers to the sender's email address within the envelope object.

Example Discovery Document

Manual Configuration Discovery is OPTIONAL. Administrators MAY configure scanner endpoints manually, including capability restrictions and authentication credentials. Manual configuration may be necessary in environments where the well-known endpoint is not accessible or where policy requires explicit configuration.

Registration Scanners register with MTAs to establish callback configuration and negotiate capabilities. Registration creates a persistent relationship that remains active until explicit deregistration or expiration.

Registration Request Scanners submit registration requests via HTTP POST to the registration endpoint. The request body contains a JSON or CBOR object with the following fields:

name: String

Human-readable name identifying the scanner.

version: String|null

Scanner software version string.

callback: CallbackConfig

Callback configuration.

expiresAt: UTCDate|null

Requested registration expiration timestamp. The MTA MAY assign a different expiration based on policy.

serialization: String

Requested serialization format for hook requests. MUST be a value from the MTA's supported serialization list.

inbound: StageSubscription|null

Inbound hook configuration. Omit or set to null if not subscribing to inbound stages.

outbound: StageSubscription|null

Outbound hook configuration. Omit or set to null if not subscribing to outbound stages.

filter: FilterOperator|FilterCondition|null

Filter configuration to limit which messages trigger hooks. Uses filter syntax from .

metadata: String[String]|null

Arbitrary key-value pairs for scanner identification and operational metadata.

A CallbackConfig object has the following properties:

url: String

HTTPS URL where the MTA sends hook requests.

timeoutMs: UnsignedInt|null

Requested timeout in milliseconds. The MTA MAY ignore or cap this value.

A StageSubscription object has the following properties:

stages: String[]

Stages to subscribe to. MUST be a subset of the MTA's supported stages for this direction.

properties: String[]|null

JSON Pointers specifying message properties to receive. A value of null requests all supported properties.

At least one of "inbound" or "outbound" MUST be present and non-null in the registration request.

Filter Configuration The filter field uses the FilterOperator and FilterCondition syntax defined in . A FilterOperator combines multiple conditions using AND, OR, or NOT logic. A FilterCondition specifies criteria that a message must match. Filter conditions follow with the following restrictions on metadata conditions: only the "size" condition from is permitted. All other metadata conditions are NOT supported as they depend on mailbox state or JMAP-specific concepts not applicable to MTA processing: inMailbox inMailboxOtherThan before after minSize maxSize allInThreadHaveKeyword someInThreadHaveKeyword noneInThreadHaveKeyword hasKeyword notKeyword All other filter conditions defined in are permitted by this specification, including text-matching conditions such as "from", "to", "cc", "bcc", "subject", "body", "header", and "text". However, support for specific conditions is determined by the MTA implementation. If a scanner specifies a filter condition that the MTA does not support, the registration MUST be rejected with an UNSUPPORTED_FILTER error. The following additional conditions are defined for envelope filtering:

envelopeFrom: String

Matches if the MAIL FROM address matches the given value. The value MAY include wildcards using the "*" character, which matches zero or more characters.

envelopeTo: String

Matches if any RCPT TO address matches the given value. The value MAY include wildcards using the "*" character, which matches zero or more characters.

Filters apply only to the "data" stage for content-based conditions (those examining message headers or body). For the "mail" and "rcpt" stages, only envelopeFrom and envelopeTo conditions are evaluated; other conditions are ignored for these stages. Messages or envelope commands not matching the filter do not trigger hook invocations.

Example Registration Request

Callback Verification Upon receiving a registration request, the MTA SHOULD verify that the callback URL is reachable and operated by the registering party. Verification proceeds as follows: The MTA generates a unique verification token. The MTA sends an HTTP POST request to the callback URL with Content-Type set to the serialization format requested in the registration request. The request body contains:

The scanner MUST respond with HTTP status 200 and a body containing the same token:

The MTA confirms the response token matches the request token. If verification fails, the MTA MUST reject the registration request: If the callback URL is unreachable or times out: HTTP 422 with error code CALLBACK_UNREACHABLE If the callback responds but the token does not match: HTTP 422 with error code CALLBACK_VERIFICATION_FAILED If the callback URL does not use HTTPS: HTTP 422 with error code TLS_REQUIRED If the callback URL's TLS certificate is invalid: HTTP 422 with error code TLS_CERTIFICATE_INVALID

Registration Response Upon successful registration, the MTA returns HTTP status 201 Created with a response body containing:

registrationId: String

Unique identifier for this registration. The identifier MUST consist of ASCII alphanumeric characters plus hyphen (-), underscore (_), and colon (:), with a maximum length of 255 characters.

status: String

Registration status. Values are "active", "suspended", or "pending_verification".

createdAt: UTCDate

Timestamp of registration creation.

expiresAt: UTCDate|null

Timestamp when registration expires, or null if no expiration is set.

callback: CallbackConfig

Confirmed callback configuration, which may differ from the request if the MTA adjusted values.

negotiated: NegotiatedCapabilities

Final negotiated capabilities representing the intersection of scanner request and MTA support.

endpoints: RegistrationEndpoints

URLs for managing this registration.

A NegotiatedCapabilities object has the following properties:

serialization: String

Confirmed serialization format.

inbound: NegotiatedSubscription|null

Confirmed inbound configuration, or null if not registered for inbound processing.

outbound: NegotiatedSubscription|null

Confirmed outbound configuration, or null if not registered for outbound processing.

A NegotiatedSubscription object has the following properties:

stages: String[]

Confirmed stages.

properties: String[]

Confirmed property list.

A RegistrationEndpoints object has the following properties:

deregistration: String

URL for deregistration requests.

status: String

URL for status queries.

Example Registration Response

Authentication Authentication for registration requests is REQUIRED but the specific mechanism is out of scope for this specification. Implementations SHOULD support at least one of the following authentication methods: Bearer tokens via the HTTP Authorization header Mutual TLS with client certificate verification HMAC-based pre-shared key authentication See for deployment guidance on authentication.

Registration Lifecycle

Status Queries Scanners MAY query their registration status via HTTP GET to the status endpoint returned during registration.

The response contains registration status and optional statistics:

registrationId: String

The registration identifier.

status: String

Current status: "active", "suspended", or "deregistered".

createdAt: UTCDate|null

Timestamp of creation.

expiresAt: UTCDate|null

Timestamp of expiration.

lastInvocation: UTCDate|null

Timestamp of most recent hook invocation.

statistics: RegistrationStatistics|null

Operational statistics.

health: HealthStatus|null

Health check status.

A RegistrationStatistics object has the following properties:

invocations: InvocationCounts

Invocation counts.

errors: InvocationCounts

Error counts with the same time period structure as invocations.

averageResponseMs: Number

Average response time in milliseconds.

An InvocationCounts object has the following properties:

total: UnsignedInt

Total count since registration.

last24h: UnsignedInt

Count in past 24 hours.

last7d: UnsignedInt

Count in past 7 days.

last30d: UnsignedInt

Count in past 30 days.

A HealthStatus object has the following properties:

status: String

Health status: "healthy", "degraded", or "unhealthy".

lastCheck: UTCDate

Timestamp of last health check.

consecutiveFailures: UnsignedInt

Count of consecutive failed health checks.

Example Status Response

Registration Expiration and Renewal Registrations MAY have an expiration time set by the MTA based on policy. Scanners MUST re-register before expiration to maintain continuous service. Re-registration follows the same process as initial registration and requires full capability renegotiation. The MTA SHOULD provide advance notice of pending expiration through the status endpoint. Scanners SHOULD monitor their registration status and initiate re-registration before expiration.

Deregistration Scanners deregister by sending an HTTP DELETE request to the deregistration endpoint.

Upon successful deregistration, the MTA returns HTTP status 200 with confirmation:

After deregistration, the MTA MUST NOT send further hook invocations to the scanner's callback URL.

In-Flight Requests Hook invocations that are in-flight at the time of deregistration MAY complete normally. The MTA SHOULD accept and process responses from these in-flight invocations. Scanners SHOULD continue to respond to hook requests that arrive during or shortly after deregistration processing, as network latency may cause requests dispatched before deregistration to arrive afterward. Once deregistration completes, the MTA MUST NOT initiate new hook invocations to the deregistered scanner's callback URL. The MTA SHOULD allow a brief grace period (implementation-defined, but typically a few seconds) for in-flight responses before considering the deregistration fully complete.

Automatic Deregistration The MTA MAY automatically deregister scanners that become unreachable. Implementation-specific policies govern the threshold for automatic deregistration, such as consecutive failure counts or timeout periods. The MTA SHOULD log automatic deregistrations for operational visibility.

Hook Request When processing reaches a stage for which scanners are registered, the MTA constructs a hook request and transmits it to each registered scanner's callback URL.

Request Structure Hook requests are transmitted via HTTP POST to the scanner's callback URL. The Content-Type header indicates the serialization format negotiated during registration. The MTA MAY include an X-MTA-Hooks-Registration header containing the registration identifier. This header is OPTIONAL but can assist scanners that handle multiple registrations at a single endpoint. The request body contains a JSON or CBOR object with properties determined by the scanner's registration. The following sections describe available properties organized by category.

Common Properties The following properties are available for both inbound and outbound processing:

stage: String

The current processing stage.

action: String

The action the MTA will perform if no scanner modifications occur. For inbound processing, values are "accept", "reject", "discard", "quarantine", or "disconnect". For outbound processing, values are "continue" or "cancel".

envelope: Envelope

The message envelope containing sender and recipient information.

message: Object|null

A structured representation of the email message based on the Email object defined in .

rawMessage: String|null

The complete message in Internet Message Format as defined in . In JSON serialization, the value is Base64-encoded. In CBOR serialization, the value is a raw byte string.

timestamp: UTCDate

Timestamp when the current stage began.

protocol: ProtocolInfo|null

Protocol information.

queue: QueueInfo|null

Queue information.

server: ServerInfo|null

Information about the MTA.

Envelope Properties An Envelope object has the following properties:

from: EnvelopeAddress

Sender information from MAIL FROM command.

to: EnvelopeAddress[]|DeliveryRecipient[]

Recipient information from RCPT TO commands. For inbound processing, this is an array of EnvelopeAddress objects. For outbound processing, this is an array of DeliveryRecipient objects.

An EnvelopeAddress object represents an address in the message envelope and has the following properties:

address: String|null

The email address. For MAIL FROM, this MAY be null to represent the null reverse-path. For RCPT TO, this MUST NOT be null.

parameters: String[String]

SMTP parameters associated with the address as key-value pairs. Common parameters include ORCPT, ENVID, and other DSN-related parameters defined in .

For outbound processing, recipient objects include additional delivery status fields described in .

Queue Properties A QueueInfo object has the following properties:

id: String

Unique queue identifier for this message within this MTA. The identifier MUST be unique for the lifetime of the message in the queue and SHOULD remain stable across hook invocations for the same queued message.

Message Properties

Structured Message The message property contains a structured representation of the email message based on the Email object defined in . The following differences from the JMAP Email object apply: Metadata properties defined in are not included, with the exception of the "size" property which MAY be present. The blobId property is replaced by a blob property containing the actual content. In JSON serialization, blob values are Base64-encoded strings. In CBOR serialization, blob values are raw byte strings. Property availability depends on the processing stage and MTA capabilities. At early stages such as "connect" or "ehlo", message properties are not available.

Raw Message The rawMessage property contains the complete message in Internet Message Format as defined in , including all MIME parts and attachments. In JSON serialization, the value is Base64-encoded. In CBOR serialization, the value is a raw byte string. The rawMessage property represents the entire message as stored or received by the MTA. When present, it provides byte-exact access to the message content, which is necessary for operations such as cryptographic signature verification. Scanners MAY request both message and rawMessage properties. If a scanner's response modifies both properties, only rawMessage modifications are applied; message modifications are ignored. The raw message content MUST NOT appear within the message property as a blob value. The message property contains only the structured parsed representation with individual body part contents available via bodyValues, while rawMessage contains the complete unparsed message.

Choosing Between Structured and Raw The message property provides a structured, parsed representation suitable for: Header inspection and modification Recipient analysis Content-type aware body part processing Efficient access to specific message components The rawMessage property provides the complete RFC 5322 message suitable for: Cryptographic verification (DKIM signature validation) Archival or compliance logging Processing that requires byte-exact message content Forwarding or re-injection without modification Scanners that need both structured access and raw content SHOULD request both properties. When modifying messages, scanners SHOULD prefer structured modifications via the message property unless byte-exact control is required. Modifications to rawMessage require the scanner to produce a complete, valid RFC 5322 message.

Server Properties A ServerInfo object has the following properties:

name: String

Server hostname.

ip: String

Server IP address.

port: UnsignedInt

Server port number.

Protocol Properties A ProtocolInfo object has the following properties:

version: String

MTA Hooks protocol version.

Inbound Properties The following properties are available only during inbound processing:

response: SmtpResponse|null

The SMTP response the MTA would send if no scanner modifications occur.

client: ClientInfo|null

Information about the connecting SMTP client.

senderAuth: SenderAuthentication|null

Results of sender authentication checks.

auth: SmtpAuthentication|null

SMTP authentication information, if the client authenticated.

tls: TlsInfo|null

TLS connection information.

SMTP Response Properties An SmtpResponse object represents an SMTP response and has the following properties:

code: UnsignedInt

The three-digit SMTP status code.

enhancedCode: String|null

The enhanced status code as defined in , if available.

message: String

The human-readable response text.

Client Properties A ClientInfo object has the following properties:

ip: String

Client IP address.

port: UnsignedInt

Client port number.

ptr: String|null

PTR record for client IP, if available.

ehlo: String

EHLO or HELO string sent by client.

asn: UnsignedInt|null

Autonomous System Number for client IP, if available.

country: String|null

ISO 3166-1 alpha-2 country code for client IP, if available.

activeConnections: UnsignedInt|null

Number of concurrent connections from this client.

TLS Properties A TlsInfo object has the following properties:

version: String

TLS protocol version (e.g., "TLSv1.3").

cipher: String

Cipher suite name.

cipherBits: UnsignedInt

Cipher key length in bits.

certIssuer: String|null

Client certificate issuer, if a client certificate was presented.

certSubject: String|null

Client certificate subject, if a client certificate was presented.

Sender Authentication Properties A SenderAuthentication object has the following properties:

spf-ehlo: String|null

SPF result for EHLO identity as defined in .

spf-mail: String|null

SPF result for MAIL FROM identity as defined in .

dkim: String|null

DKIM verification result as defined in .

arc: String|null

ARC verification result.

dmarc: String|null

DMARC evaluation result as defined in .

Authentication result values follow the conventions established in the respective specifications.

SMTP Authentication Properties An SmtpAuthentication object has the following properties:

login: String

Authenticated username.

method: String

SASL mechanism used for authentication.

Outbound Properties The following properties are available only during outbound processing.

Extended Queue Properties For outbound processing, the queue property uses an OutboundQueueInfo object, which extends QueueInfo with additional fields:

id: String

Unique queue identifier for this message.

expiresAt: UTCDate|null

Timestamp when the queue entry expires.

attempts: UnsignedInt

Total delivery attempts made so far.

Recipient Delivery Status For outbound processing, each recipient in envelope.to is represented as a DeliveryRecipient object. A DeliveryRecipient object extends EnvelopeAddress with delivery status information and has the following properties:

address: String

The recipient email address.

parameters: String[String]

SMTP parameters associated with the address as key-value pairs.

status: String

The delivery status for this recipient. Values are "pending", "delivered", "deferred", "failed", or "failed-silent".

attempt: UnsignedInt

The current delivery attempt number for this recipient.

lastResponse: SmtpResponse|null

The last SMTP response received for this recipient, or null if no attempt has been made.

nextAttemptAt: UTCDate|null

The scheduled time for the next delivery attempt, or null if no retry is scheduled.

nextDsnAt: UTCDate|null

The scheduled time for the next DSN generation, or null if no DSN is scheduled.

queueName: String|null

The name of the outbound queue handling this recipient, if applicable.

For the "dsn" stage, the message and rawMessage properties contain the DSN message that the MTA is about to send. Scanners MAY modify or delete these properties to alter or suppress DSN generation.

Example Inbound Hook Request

"], "headers": [ {"name": "From", "value": "Sender Name "}, {"name": "To", "value": "recipient@example.com"}, {"name": "Subject", "value": "Meeting Tomorrow"}, {"name": "Date", "value": "Sat, 21 Dec 2024 16:29:00 +0000"}, {"name": "Message-ID", "value": ""} ], "bodyStructure": { "type": "text/plain", "charset": "utf-8", "size": 28 }, "bodyValues": { "1": { "value": "Let's meet tomorrow at 10am.", "isEncodingProblem": false, "isTruncated": false } } }, "client": { "ip": "192.0.2.100", "port": 54321, "ptr": "mail.example.org", "ehlo": "mail.example.org", "activeConnections": 3 }, "server": { "name": "mx1.example.com", "ip": "198.51.100.25", "port": 25 }, "tls": { "version": "TLSv1.3", "cipher": "TLS_AES_256_GCM_SHA384", "cipherBits": 256 }, "senderAuth": { "spf-mail": "pass", "dkim": "pass", "dmarc": "pass" } } ]]>

Example Outbound Hook Request

Hook Response Scanners respond to hook requests with modifications to the request object. These modifications instruct the MTA how to proceed with message processing.

Response Structure Scanners respond with an HTTP 200 status code. The response body contains a JSON or CBOR object specifying modifications to apply to the hook request. The object has the following fields:

set: SetOperation[]|null

An array of set operations to apply, or null if no set operations.

add: AddOperation[]|null

An array of add operations to apply, or null if no add operations.

delete: DeleteOperation[]|null

An array of delete operations to apply, or null if no delete operations.

A SetOperation object has the following properties:

path: String

JSON Pointer (per ) to the property to replace.

value: any

The new value for the property.

An AddOperation object has the following properties:

path: String

JSON Pointer to the location where the value should be added.

value: any

The value to add.

index: UnsignedInt|null

For array additions, the zero-based index at which to insert the value. If null or omitted, the value is appended to the array.

A DeleteOperation object has the following properties:

path: String

JSON Pointer to the property to remove.

Actions Scanners change the MTA's action by including a set operation targeting the "/action" path. The following values are valid for inbound and outbound processing respectively. The following action values are valid:

Inbound Actions

accept

Accept the message for delivery to recipients. Non-terminal; chain continues.

reject

Reject the message and return an error response to the sending client. Terminal; chain stops.

discard

Accept the message from the client but do not deliver it. The client receives a success response. Terminal; chain stops.

quarantine

Accept the message and place it in quarantine for administrative review. Quarantine location and handling are implementation-defined. Non-terminal; chain continues.

disconnect

Terminate the SMTP connection immediately. Terminal; chain stops.

Outbound Actions

continue

Proceed with normal processing. Non-terminal; chain continues.

cancel

Cancel all pending deliveries for this message. Terminal; chain stops.

Modifications Scanners communicate changes by specifying operations on the hook request object. The response contains JSON Pointer paths identifying properties to modify and the operations to perform. These modifications can alter any aspect of the transaction that the MTA permits, including: The action the MTA should take (accept, reject, quarantine, etc.) The SMTP response to send to the client or expect from the server Message headers and body content Envelope addresses (sender and recipients) Per-recipient delivery status for outbound processing Three modification operations are supported:

set

Replace the value at a specified path with a new value.

add

Insert a value at a specified path. For objects, this adds a new property. For arrays, this inserts an element at the specified index or appends if no index is given.

delete

Remove the value at a specified path.

Set Operations Set operations replace values at specified paths.

Add Operations Add operations insert new values. For arrays, an optional index specifies insertion position.

Delete Operations Delete operations remove values at specified paths. When deleting array elements, indices refer to positions in the array at the time each delete operation executes. Because delete operations execute sequentially and earlier deletions cause subsequent elements to shift to lower indices, scanners SHOULD order array deletions from highest to lowest index to avoid unexpected results. For example, to delete elements originally at indices 1 and 3 from an array:

Deleting index 3 first leaves the element originally at index 1 still at index 1. If the order were reversed, deleting index 1 first would shift the element originally at index 3 to index 2.

Modification Order The MTA applies modifications in the following order: Set operations Add operations Delete operations This ordering allows predictable results when multiple operations affect related paths. To change the action the MTA takes, the scanner sets the "/action" path to the desired value. To modify the SMTP response, the scanner can either set individual response fields (e.g., "/response/code", "/response/message") or replace the entire response object by setting "/response". If a scanner modifies both the "message" and "rawMessage" properties, the "rawMessage" modification takes precedence and "message" modifications are ignored.

Conflicting Operations If a response contains multiple operations targeting the same path or overlapping paths, the MTA applies them in the defined order (set, then add, then delete). The final state reflects all operations applied sequentially. For example, if a response sets "/message/subject" and also deletes "/message/subject", the subject will be deleted (delete operations execute last). Implementations SHOULD NOT submit responses with conflicting operations targeting the same path, as the resulting behavior, while deterministic, may be confusing.

No Modification Response If a scanner has no modifications to request, it MUST return either: An empty JSON/CBOR object: {} A response with empty or null modification arrays: {"set": null, "add": null, "delete": null} An HTTP 204 No Content response with no body In all cases, the MTA proceeds with the action specified in the original request. The scanner chain continues to the next registered scanner unless the current action is terminal.

Size Limits MTAs MAY impose limits on the size of modification values. If a modification would cause a header or message to exceed configured size limits, the MTA SHOULD reject that specific modification and log the failure.

Modification Errors If a modification cannot be applied (for example, the path references a non-existent property for deletion, or the path is not in the permitted updateProperties list), the MTA SHOULD: Log the error with sufficient detail for debugging Increment modification failure statistics Continue processing remaining modifications in the response The MTA MUST NOT reject the entire response due to a single failed modification unless the failed modification is critical to interpreting the response (such as an invalid action value).

Example Hook Responses

Accept with Header Addition

Reject with Custom Response

Modify Subject and Add Recipient

Cancel Specific Recipient Delivery For outbound processing, to cancel delivery for a specific recipient:

No Action Response If a scanner has no changes to request, it returns an empty response body or a response with no action or modifications:

The MTA proceeds with the default action specified in the request.

Transport This section specifies HTTP transport requirements for MTA Hooks communications.

HTTP Version Implementations MUST support HTTP/1.1 . Implementations SHOULD support HTTP/2 for improved performance through multiplexing. Implementations MAY support HTTP/3 . When multiple HTTP versions are available, implementations SHOULD prefer newer versions for their performance benefits while maintaining fallback to HTTP/1.1.

TLS Requirements All MTA Hooks communications MUST use TLS. Implementations MUST support TLS 1.2 and SHOULD support TLS 1.3 . Implementations MUST validate server certificates against trusted certificate authorities. Self-signed certificates SHOULD NOT be accepted in production deployments unless explicitly configured by administrators. Cipher suite selection SHOULD follow current best practices. Implementations SHOULD disable cipher suites known to be weak or compromised.

Request Methods The following HTTP methods are used: GET: Discovery document retrieval and status queries POST: Registration requests and hook invocations DELETE: Deregistration requests

Content Negotiation For hook requests and responses, the Content-Type header MUST be set to "application/json" for JSON serialization or "application/cbor" for CBOR serialization. The Accept header MAY be used during discovery to indicate preferred serialization format. If the scanner cannot provide the requested format, it SHOULD return the discovery document in JSON format.

Error Handling

HTTP Status Codes MTAs SHOULD interpret HTTP status codes as follows: 2xx: Request successful. Process response body. 4xx: Client error. Do not retry; log error and proceed with default action. 5xx: Server error. May retry according to policy.

Timeout Handling If a scanner does not respond within the configured timeout, the MTA SHOULD proceed with the default action. Timeout handling policy (fail-open vs. fail-closed) is implementation-defined and SHOULD be configurable.

Retry Policy For transient errors (5xx status codes, network failures, timeouts), implementations SHOULD implement retry with exponential backoff. A recommended policy is: Maximum 3 retry attempts Initial delay: 100 milliseconds Backoff multiplier: 2 Maximum delay: 5 seconds Add random jitter of 0-100 milliseconds Implementations MAY provide configuration options to adjust retry parameters.

Connection Management Implementations SHOULD use connection pooling to reduce latency for repeated requests to the same scanner endpoint. HTTP/2 multiplexing, where available, reduces the need for multiple connections. Implementations SHOULD respect HTTP keep-alive semantics and connection limits advertised by scanners.

Implementation Considerations

MTA Considerations

Multiple Scanner Handling When multiple scanners are registered for the same stage, the MTA invokes them sequentially in an implementation-defined order. Each scanner's response affects the request seen by subsequent scanners. Implementations SHOULD provide configuration options for: Scanner invocation order (priority-based or explicit ordering) Conflict resolution when scanners return conflicting actions Short-circuit behavior (whether to skip remaining scanners after certain actions)

Unavailable Scanners When a scanner becomes unavailable (timeouts, errors, deregistration), the MTA must determine how to proceed. Common policies include: Fail-open: Proceed with default action as if scanner approved Fail-closed: Reject or defer message processing The appropriate policy depends on deployment requirements. Security-focused deployments may prefer fail-closed, while availability-focused deployments may prefer fail-open. Implementations SHOULD make this policy configurable.

Performance Considerations Synchronous hook invocation adds latency to mail processing. Implementations SHOULD consider: Connection pooling and keep-alive for scanner connections Parallel invocation of independent scanners where ordering is not required Timeout tuning based on scanner response time characteristics

Scanner Considerations

Idempotency Scanners SHOULD design their processing to be idempotent. Network issues or MTA retries may result in duplicate invocations for the same message. Scanners SHOULD handle duplicate requests gracefully.

Response Time Scanners operate in the critical path of mail delivery. Long response times delay message processing and may cause timeouts. Scanners SHOULD: Process requests promptly Implement internal timeouts shorter than MTA timeouts Consider asynchronous processing for expensive operations Return default responses when processing cannot complete in time

Stateless Design Scanners SHOULD avoid relying on state from previous invocations. Each hook request contains complete context for processing decisions. Stateless design improves reliability and simplifies horizontal scaling.

Large Message Handling Large messages present challenges for both MTAs and scanners. Implementations SHOULD consider: Message size limits advertised in discovery documents Truncation policies for oversized messages Memory management for large message bodies Streaming or chunked delivery is not currently specified. For very large messages, implementations MAY arrange out-of-band content retrieval, though this is outside the scope of this specification.

High Availability

Scanner High Availability For production deployments, scanners SHOULD be deployed with redundancy. Approaches include: Multiple scanner instances behind a load balancer Health checking with automatic failover Geographic distribution for disaster recovery The MTA registers a single callback URL; load balancing occurs at the network layer.

Registration State Registration state resides at the MTA. Scanner instances SHOULD NOT assume persistent local state. If scanner instances share a registration, they MUST coordinate to avoid conflicting operations (such as simultaneous deregistration).

Security Considerations

Authentication and Authorization Authentication of registration requests is critical to prevent unauthorized scanners from receiving email content. Implementations MUST require authentication for registration operations. Recommended authentication approaches: Bearer tokens: Simple to implement; tokens should be generated with sufficient entropy and rotated periodically. Mutual TLS: Provides strong authentication and binds identity to transport security. HMAC signatures: Allows authentication without transmitting secrets; requires secure key distribution. Authorization policies SHOULD restrict which scanners may register and what capabilities they may request. Policies may be based on scanner identity, requested stages, or other criteria.

Transport Security All MTA Hooks communications MUST use TLS to protect message content and authentication credentials in transit. Implementations MUST validate certificates to prevent man-in-the-middle attacks. For internal network deployments, administrators may choose to use private certificate authorities. However, disabling certificate validation is NOT RECOMMENDED even for internal communications.

Message Confidentiality Email messages frequently contain sensitive personal or business information. Scanners necessarily receive access to message content to perform their functions. Deployments SHOULD consider: Data handling policies for scanner operators Logging policies that avoid recording message content Data retention limits at scanner endpoints Regulatory requirements such as GDPR for personal data processing

Denial of Service

Registration Attacks Attackers may attempt to exhaust MTA resources through excessive registration requests. Implementations SHOULD: Rate limit registration attempts per source Limit maximum concurrent registrations Require authentication before processing registrations Validate callback URLs before completing registration

Scanner Resource Exhaustion Malicious or misconfigured MTAs could overwhelm scanners with requests. Scanners SHOULD: Implement rate limiting Set appropriate resource limits Monitor for unusual request patterns

Slow Scanner Attacks Slow responses from scanners can exhaust MTA resources (connection limits, memory). MTAs SHOULD: Enforce strict timeouts Limit concurrent requests to any single scanner Implement circuit breaker patterns for repeatedly slow scanners

Injection Attacks Scanner responses contain modifications applied to message processing. Implementations MUST validate all scanner-provided data: Header values MUST be validated to prevent header injection Recipient addresses MUST be validated against policy Body modifications MUST not introduce malformed content

Privacy Considerations Email processing inherently involves access to personal communications. Implementations SHOULD minimize data exposure: Scanners SHOULD request only necessary properties MTAs SHOULD honor property restrictions from capability negotiation Logging SHOULD avoid recording message content or metadata Scanner operators SHOULD implement appropriate data protection measures

IANA Considerations

Well-Known URI Registration This document registers the following well-known URI per : URI suffix: mta-hooks Change controller: IETF Specification document: This document Status: permanent Related information: N/A

MTA Hooks Serialization Format Registry IANA is requested to create a new registry entitled "MTA Hooks Serialization Formats" with the following initial contents: Format: json Description: JSON serialization per RFC 8259 Reference: This document Format: cbor Description: CBOR serialization per RFC 8949 Reference: This document New registrations require Specification Required per .

MTA Hooks Inbound Action Registry IANA is requested to create a new registry entitled "MTA Hooks Inbound Actions" with the following initial contents: Action: accept Description: Accept message for delivery Reference: This document Action: reject Description: Reject message with error response Reference: This document Action: discard Description: Accept but do not deliver message Reference: This document Action: quarantine Description: Place message in quarantine Reference: This document Action: disconnect Description: Terminate SMTP connection Reference: This document New registrations require Specification Required per .

MTA Hooks Outbound Action Registry IANA is requested to create a new registry entitled "MTA Hooks Outbound Actions" with the following initial contents: Action: continue Description: Proceed with normal delivery processing Reference: This document Action: cancel Description: Cancel pending deliveries Reference: This document New registrations require Specification Required per .

MTA Hooks Inbound Stage Registry IANA is requested to create a new registry entitled "MTA Hooks Inbound Stages" with the following initial contents: Stage: connect Description: Client connection established Reference: This document Stage: ehlo Description: EHLO/HELO command received Reference: This document Stage: mail Description: MAIL FROM command received Reference: This document Stage: rcpt Description: RCPT TO command received Reference: This document Stage: data Description: Message content received Reference: This document New registrations require Specification Required per .

MTA Hooks Outbound Stage Registry IANA is requested to create a new registry entitled "MTA Hooks Outbound Stages" with the following initial contents: Stage: delivery Description: Delivery attempt completed Reference: This document Stage: defer Description: Delivery deferred for retry Reference: This document Stage: dsn Description: DSN generation pending Reference: This document New registrations require Specification Required per .

MTA Hooks Error Code Registry IANA is requested to create a new registry entitled "MTA Hooks Error Codes" with the initial contents specified in . New registrations require Specification Required per .

This document is a specification of the basic protocol for Internet electronic mail transport. It consolidates, updates, and clarifies several previous documents, making all or parts of most of them obsolete. It covers the SMTP extension mechanisms and best practices for the contemporary Internet, but does not provide details about particular extensions. Although SMTP was designed as a mail transport and delivery protocol, this specification also contains information that is important to its use as a "mail submission" protocol for "split-UA" (User Agent) mail reading systems and mobile environments. [STANDARDS-TRACK] This document specifies the Internet Message Format (IMF), a syntax for text messages that are sent between computer users, within the framework of "electronic mail" messages. This specification is a revision of Request For Comments (RFC) 2822, which itself superseded Request For Comments (RFC) 822, "Standard for the Format of ARPA Internet Text Messages", updating it to reflect current practice and incorporating incremental changes that were specified in other RFCs. [STANDARDS-TRACK] JSON Pointer defines a string syntax for identifying a specific value within a JavaScript Object Notation (JSON) document. JavaScript Object Notation (JSON) is a lightweight, text-based, language-independent data interchange format. It was derived from the ECMAScript Programming Language Standard. JSON defines a small set of formatting rules for the portable representation of structured data. This document removes inconsistencies with other specifications of JSON, repairs specification errors, and offers experience-based interoperability guidance. This document specifies a protocol for clients to efficiently query, fetch, and modify JSON-based data objects, with support for push notification of changes and fast resynchronisation and for out-of- band binary data upload/download. This document specifies a data model for synchronising email data with a server using the JSON Meta Application Protocol (JMAP). Clients can use this to efficiently search, access, organise, and send messages, and to get push notifications for fast resynchronisation when new messages are delivered or a change is made in another client. The Concise Binary Object Representation (CBOR) is a data format whose design goals include the possibility of extremely small code size, fairly small message size, and extensibility without the need for version negotiation. These design goals make it different from earlier binary serializations such as ASN.1 and MessagePack. This document obsoletes RFC 7049, providing editorial improvements, new details, and errata fixes while keeping full compatibility with the interchange format of RFC 7049. It does not create a new version of the format. This document defines a date and time format for use in Internet protocols that is a profile of the ISO 8601 standard for representation of dates and times using the Gregorian calendar. In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings. The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document specifies the HTTP/1.1 message syntax, message parsing, connection management, and related security concerns. This document obsoletes portions of RFC 7230. This specification describes an optimized expression of the semantics of the Hypertext Transfer Protocol (HTTP), referred to as HTTP version 2 (HTTP/2). HTTP/2 enables a more efficient use of network resources and a reduced latency by introducing field compression and allowing multiple concurrent exchanges on the same connection. This document obsoletes RFCs 7540 and 8740. The QUIC transport protocol has several features that are desirable in a transport for HTTP, such as stream multiplexing, per-stream flow control, and low-latency connection establishment. This document describes a mapping of HTTP semantics over QUIC. This document also identifies HTTP/2 features that are subsumed by QUIC and describes how HTTP/2 extensions can be ported to HTTP/3. This document specifies Version 1.2 of the Transport Layer Security (TLS) protocol. The TLS protocol provides communications security over the Internet. The protocol allows client/server applications to communicate in a way that is designed to prevent eavesdropping, tampering, or message forgery. [STANDARDS-TRACK] This document specifies version 1.3 of the Transport Layer Security (TLS) protocol. TLS allows client/server applications to communicate over the Internet in a way that is designed to prevent eavesdropping, tampering, and message forgery. This document updates RFCs 5705 and 6066, and obsoletes RFCs 5077, 5246, and 6961. This document also specifies new requirements for TLS 1.2 implementations. This memo defines a path prefix for "well-known locations", "/.well-known/", in selected Uniform Resource Identifier (URI) schemes. In doing so, it obsoletes RFC 5785 and updates the URI schemes defined in RFC 7230 to reserve that space. It also updates RFC 7595 to track URI schemes that support well-known URIs in their registry. Many protocols make use of points of extensibility that use constants to identify various protocol parameters. To ensure that the values in these fields do not have conflicting uses and to promote interoperability, their allocations are often coordinated by a central record keeper. For IETF protocols, that role is filled by the Internet Assigned Numbers Authority (IANA). To make assignments in a given registry prudently, guidance describing the conditions under which new values should be assigned, as well as when and how modifications to existing values can be made, is needed. This document defines a framework for the documentation of these guidelines by specification authors, in order to assure that the provided guidance for the IANA Considerations is clear and addresses the various issues that are likely in the operation of a registry. This is the third edition of this document; it obsoletes RFC 5226. This memo defines an extension to the Simple Mail Transfer Protocol (SMTP) service, which allows an SMTP client to specify (a) that Delivery Status Notifications (DSNs) should be generated under certain conditions, (b) whether such notifications should return the contents of the message, and (c) additional information, to be returned with a DSN, that allows the sender to identify both the recipient(s) for which the DSN was issued, and the transaction in which the original message was sent. [STANDARDS-TRACK] DomainKeys Identified Mail (DKIM) permits a person, role, or organization that owns the signing domain to claim some responsibility for a message by associating the domain with the message. This can be an author's organization, an operational relay, or one of their agents. DKIM separates the question of the identity of the Signer of the message from the purported author of the message. Assertion of responsibility is validated through a cryptographic signature and by querying the Signer's domain directly to retrieve the appropriate public key. Message transit from author to recipient is through relays that typically make no substantive change to the message content and thus preserve the DKIM signature. This memo obsoletes RFC 4871 and RFC 5672. [STANDARDS-TRACK] Email on the Internet can be forged in a number of ways. In particular, existing protocols place no restriction on what a sending host can use as the "MAIL FROM" of a message or the domain given on the SMTP HELO/EHLO commands. This document describes version 1 of the Sender Policy Framework (SPF) protocol, whereby ADministrative Management Domains (ADMDs) can explicitly authorize the hosts that are allowed to use their domain names, and a receiving host can check such authorization. This document obsoletes RFC 4408. Domain-based Message Authentication, Reporting, and Conformance (DMARC) is a scalable mechanism by which a mail-originating organization can express domain-level policies and preferences for message validation, disposition, and reporting, that a mail-receiving organization can use to improve mail handling. Originators of Internet Mail need to be able to associate reliable and authenticated domain identifiers with messages, communicate policies about messages that use those identifiers, and report about mail using those identifiers. These abilities have several benefits: Receivers can provide feedback to Domain Owners about the use of their domains; this feedback can provide valuable insight about the management of internal operations and the presence of external domain name abuse. DMARC does not produce or encourage elevated delivery privilege of authenticated email. DMARC is a mechanism for policy distribution that enables increasingly strict handling of messages that fail authentication checks, ranging from no action, through altered delivery, up to message rejection.

Error Codes This appendix defines error codes used in MTA Hooks error responses.

Error Response Structure Error responses use the following structure:

HTTP 400 Bad Request INVALID_REQUEST: Malformed JSON/CBOR or missing required fields. INVALID_CALLBACK_URL: Callback URL is malformed or uses disallowed scheme. CAPABILITY_MISMATCH: Requested capability not supported by MTA. INVALID_STAGE: One or more requested stages are not valid. INVALID_ACTION: One or more requested actions are not valid. INVALID_MODIFICATION: One or more requested modifications are not valid. NO_STAGES_REQUESTED: Neither inbound nor outbound stages were specified. FILTER_INVALID: Filter configuration is syntactically invalid.

HTTP 401 Unauthorized AUTHENTICATION_REQUIRED: No authentication credentials provided. INVALID_CREDENTIALS: Provided credentials are invalid or expired.

HTTP 403 Forbidden REGISTRATION_DENIED: Valid credentials but not authorized to register. DOMAIN_NOT_ALLOWED: Scanner not authorized for requested domains. CALLBACK_NOT_ALLOWED: Callback URL not in allowlist.

HTTP 404 Not Found REGISTRATION_NOT_FOUND: No registration exists with the specified ID.

HTTP 409 Conflict ALREADY_REGISTERED: Scanner with same callback URL already registered. REGISTRATION_LIMIT_REACHED: Maximum number of registrations exceeded.

HTTP 422 Unprocessable Entity UNSUPPORTED_FILTER: The filter contains conditions that are not supported by this MTA. CALLBACK_UNREACHABLE: MTA could not reach callback URL for verification. CALLBACK_VERIFICATION_FAILED: Callback URL responded but verification failed. TLS_REQUIRED: Callback URL must use HTTPS. TLS_CERTIFICATE_INVALID: Callback URL TLS certificate is invalid or untrusted.

HTTP 429 Too Many Requests RATE_LIMITED: Too many registration attempts; retry after specified time.

HTTP 500 Internal Server Error INTERNAL_ERROR: Unexpected server error during processing.

HTTP 503 Service Unavailable REGISTRATION_DISABLED: Registration temporarily disabled by administrator.

Complete Example Flows This appendix provides complete request and response examples for common scenarios.

Inbound Spam Filtering This example shows a complete flow for spam filtering during inbound message reception.

Discovery

Registration

Verification Callback

Registration Response

Hook Invocation - Clean Message

Hook Invocation - Spam Detected

Outbound Delivery Logging This example shows outbound hook usage for delivery status logging and compliance.

Registration

Hook Invocation - Partial Delivery

", "size": 2048 } } ]]>

Deregistration

Acknowledgments The authors thank the developers and operators of existing mail filtering systems whose experience informed this design.

Changes [[This section to be removed by RFC Editor]] draft-degennaro-mta-hooks-00 Initial version