Many spam and abuse prevention techniques rely on users being able to report other users who are sending unwanted messages, or specific instances of abuse. &xep0191; allows users to block spammers, but does not provide a mechanism for them to report a reason for the block to the server operator. This specification extends the blocking command to optionally provide an abuse report.

This document extends the blocking command instead of providing a separate reporting IQ because we hypothesize that this will slightly lower the levels of false reports received by service operators. We have observed a common pattern on the internet where a user becomes mad at or disagrees with another user and begins harassing them by replying to or reporting their every comment even if it is not itself spam or abusive. However, this sort of behavior cannot continue if the harasser can no longer read the messages of the person they are stalking. Giving them a choice between their abusive behavior and being able to read their targets can possibly force them to break the cycle and only create valid reports.

Entities that support &xep0030; and abuse reporting using the blocking command as defined in this spec MUST respond to service discovery requests with a feature of 'urn:xmpp:reporting:1'. Support for this namespace also indicates support for the abuse reporting reasons defined in this document. For example, a response from a server that supports reporting and understands the abuse and spam reasons defined later in this specification might look like the following:

The payload for reporting abuse to the server takes the form of a <report/> qualified by the 'urn:xmpp:reporting:1' namespace &VNOTE;.

Abuse reports MUST include a reason for the report in the "reason" attribute.

This document defines the following reasons for a report:

urn:xmpp:reporting:spam

Used for reporting a JID that is sending unwanted messages.

urn:xmpp:reporting:abuse

Used for reporting general abuse.

Reports MAY contain a user provided message explaining or providing context about the reason for the report. See also the Internationalization Considerations section of this document.

To send a report, a report payload MAY be inserted into an <item/> node sent as part of a request to block a spammer as defined in &xep0191;. For example:

Servers that receive a blocking command with a report MUST block the JID or return an error just as they would if no report were present. Servers then MAY take other actions based on the report, however, such actions are outside the scope of this document.

If the server supports &xep0313; the report MAY also include the stanza-id of specific messages being reported. This is done by including copies of each <stanza-id/> element that the user wishes to report as a child of the <report/> element. The stanza indicated by the provided stanza-id SHOULD be by the same JID being reported and blocked.

Clients that support sending reports as part of the blocking command SHOULD expose interfaces to both block a JID without reporting it as abuse, and to block and report a JID.

The blocking command may be used to block multiple JIDs at the same time. When blocking multiple JIDs any abuse report only applies to a single JID. If the client allows selecting multiple JIDs in an abuse reporting dialog they SHOULD also allow choosing a separate reason, text, and messages for each JID. They MAY choose to only allow reporting a single JID at a time as well when the "block and report" dialog is accessed, and multiple JIDs when the "block" dialog is accessed.

If one or more <text/> elements are present they SHOULD include 'xml:lang' attributes specifying the natural language of the XML character data.

This document introduces no additional security considerations above and beyond those defined in the documents on which it depends.

This document requires no interaction with &IANA;.

An XML schema will be added before this specification moves to draft status.

Thanks to the participants of the XMPP Summit 20 in Austin, TX who discussed this XEP: specifically to Waqas Hussain, Kevin Smith, Lance Stout, and Matthew Wild. A special thanks to Daniel Wisnewski for giving the presentation that kicked off the anti-abuse work.

Thanks also (in no particular order) to Jonas Wielicki, Georg Lukas, Daniel Gultsch, and Matthew Wild for their feedback.