There are various spim protection methods exist in XMPP: Privacy Lists (XEP-0016) , CAPTCHA Forms (XEP-0158) , Blocking Command (XEP-0191) , Incident Reporting (XEP-0268)  and Entity Reputation (XEP-0275) . But they may not be sufficient enough:
Service administrators might want to deploy server-based spim recognition software to fill in the gaps. However, every automated spim recognition suffers from false positives - situations where a stanza incorrectly qualified as spim. To avoid them, a spim filter doesn't block suspicious stanza, but marks it and sends to a client in a regular manner. A client software doesn't need to interrupt a user when processing such marked stanzas: for example, it may put them silently in "SPAM" folder, so a user can look through them at any time later. Furthermore, a spim filter may take user's experience into account. When a user receives an unsolicited stanza, he or she can mark it as spim. In this case a client software sends an automatic complaint to a server-based spim filter. This specification deals with both cases. Thus, in contrast to SPIM-Blocking Control (XEP-0159) , it doesn't introduce any spim blocking techniques. Also, the various spim recognition procedures that may be employed by the server are beyond the scope of this document.
An implementation compliant with this document MUST support spim markers as described in Spim Marker use case. Support for spim reports, as described in Spim Report use case, is RECOMMENDED.
The following terms are used throughout this document:
The filtering entity marks abusive stanza by adding <mark/> child element qualified by the 'urn:xmpp:spim-marker:0' namespace. The element MUST possess the 'filter' attribute whose value MUST be a full jid of the filtering entity. The <mark/> element MAY contain character data which SHOULD be a human-readable description of the reason to mark. The filtering entity MUST NOT add more than one <mark/> element and MUST delete all other <mark/> elements matching itself before adding a new one. The filtering entity MAY remove any <mark/> elements matching itself even if it doesn't add a new one.
Processing rules of marked stanzas taken by the receiving entity are beyond the scope of this document. One possible solution is to put such stanzas silently in so-called "SPAM" folder.
If the filtering entity wishes to receive abuse report for the stanza, it MUST add <report/> child element qualified by the 'urn:xmpp:spim-report:0' namespace and MUST possess the 'key' and the 'filter' attributes. A value of the 'key' attribute is arbitrary, but SHOULD have at least 128 bits of randomness. The 'key' attribute is needed to match the corresponding complaint (if any) with the sender. The value of the 'filter' attribute MUST be a full jid of the filtering entity. The filtering entity MUST NOT add more than one <report/> element and MUST delete all other <report/> elements matching itself before adding a new one. The filtering entity MAY remove any <report/> elements matching itself even if it doesn't add a new one.
The receiving entity MAY complain by sending an IQ-set containing the <query/> child element qualified by the 'urn:xmpp:spim-report:0' namespace. A value of the 'filter' attribute MUST be copied in the 'to' attribute of the IQ-set stanza. The element MUST possess 'key' attribute copied from the original stanza.
The receiving entity MUST ignore any <report/> elements generated by untrusted filtering entities. If there are more than one <report/> element matching the same filtering entity, all of them MUST be ignored.
The filtering entity MUST respond with an empty IQ-result stanza upon successful completion of the request:
A filtering entity SHOULD only add <mark/> or <report/> elements and a receiving entity SHOULD only process those elements if the corresponding stanza envolves an interaction with a human user: subscription requests, messages, conference invites, voice calls, etc. For example, it doesn't make a lot of sense to mark Software Information (XEP-0232)  stanzas.
To avoid obvious false positives and user confusions, a filtering entity SHOULD NOT add <mark/> or <report/> elements to a stanza and a receiving entity SHOULD ignore <mark/> and <report/> elements of a stanza if:
If an entity supports the spim markers, it MUST report that by including a service discovery feature of "urn:xmpp:spim-marker:0" in response to a Service Discovery (XEP-0030)  information request. If an entity supports the spim reports, it MUST report that by including a service discovery feature of "urn:xmpp:spim-report:0" in response to a Service Discovery (XEP-0030)  information request:
Care should be taken if a receiving entity chooses to generate a CAPTCHA challenge (CAPTCHA Forms (XEP-0158) ) in response to a marked stanza. A spim recognition system rarely has more than 5-10% of false positives. Thus, producing CAPTCHA images or audio/video samples is likely a waste of system resources and also may overload the receiving entity at high rate of spim stanzas.
A rogue server may add fake <mark/> elements to compromise filtering entities: a user may decide to remove such entities from the trusted list because, for example, he or she thinks they produce too many false positives. To avoid such situation, a filtering entity MUST remove any <mark/> elements matching itself before adding new <mark/> element as described in Spim Marker use case. Also, a filtering entity MAY remove any <mark/> elements matching itself even if it doesn't add a new one.
An attacker may add fake <report/> element. For example, it may do that for checking an activity of the user. To avoid such situation, a receiving entity MUST send spim reports to the trusted filtering entities only as desribed in Spim Report use case.
An attacker may add thousands of fake <report/> elements matching the single trusted filtering entity in one stanza. A poorly written receiving entity may generate a complaint for all of them. As an effect, a distributed DoS attack on the filtering entity is performed if there are multiple receiving entities envolved. To avoid such situation, a receiving entity MUST ignore multiple <report/> elements matching the same filtering entity as desribed in Spim Report use case.
In its turn, a filtering entity MUST remove any <report/> elements matching itself before adding new <report/> element as described in Spim Report use case. Thus, it is guaranteed that the element will not be ignored by the receiving entity.
An attacker may gain an information about user's trusted filtering entities. In this case he or she may add the <report/> element per every such entity in one stanza. If there are too many filtering entities in the list, a user may generate enormous traffic when generating spim reports. Although this attack is not very effective, a client software MUST not generate spim reports without user's acknowledgement.
An attacker may try to mark an innocent user as a spimmer by producing several IQ-set stanzas qualified by "urn:xmpp:spim-report:0" containing different value of the 'key' attribute each (so-called "dictionary attack"). As a protection, sanity checks MUST be performed when processing such reports. For example, if a filtering entity doesn't store any information about a receiving entity, the value of the 'key' attribute SHOULD have at least 128 bits of randomness.
This document requires no interaction with the Internet Assigned Numbers Authority (IANA) .
This specification defines the following XML namespaces:
Upon advancement of this specification from a status of Experimental to a status of Draft, the XMPP Registrar  shall add the foregoing namespace to the registry located at <https://xmpp.org/registrar/namespaces.html>, as described in Section 4 of XMPP Registrar Function (XEP-0053) .
Thanks to Sergei Golovan for the feedback.
This document in other formats: XML PDF
This XMPP Extension Protocol is copyright © 1999 – 2020 by the XMPP Standards Foundation (XSF).
Permission is hereby granted, free of charge, to any person obtaining a copy of this specification (the "Specification"), to make use of the Specification without restriction, including without limitation the rights to implement the Specification in a software program, deploy the Specification in a network service, and copy, modify, merge, publish, translate, distribute, sublicense, or sell copies of the Specification, and to permit persons to whom the Specification is furnished to do so, subject to the condition that the foregoing copyright notice and this permission notice shall be included in all copies or substantial portions of the Specification. Unless separate permission is granted, modified works that are redistributed shall not contain misleading information regarding the authors, title, number, or publisher of the Specification, and shall not claim endorsement of the modified works by the authors, any organization or project to which the authors belong, or the XMPP Standards Foundation.
## NOTE WELL: This Specification is provided on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. ##
In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall the XMPP Standards Foundation or any author of this Specification be liable for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising from, out of, or in connection with the Specification or the implementation, deployment, or other use of the Specification (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if the XMPP Standards Foundation or such author has been advised of the possibility of such damages.
This XMPP Extension Protocol has been contributed in full conformance with the XSF's Intellectual Property Rights Policy (a copy of which can be found at <https://xmpp.org/about/xsf/ipr-policy> or obtained by writing to XMPP Standards Foundation, P.O. Box 787, Parker, CO 80134 USA).
The Extensible Messaging and Presence Protocol (XMPP) is defined in the XMPP Core (RFC 6120) and XMPP IM (RFC 6121) specifications contributed by the XMPP Standards Foundation to the Internet Standards Process, which is managed by the Internet Engineering Task Force in accordance with RFC 2026. Any protocol defined in this document has been developed outside the Internet Standards Process and is to be understood as an extension to XMPP rather than as an evolution, development, or modification of XMPP itself.
The primary venue for discussion of XMPP Extension Protocols is the <email@example.com> discussion list.
Discussion on other xmpp.org discussion lists might also be appropriate; see <http://xmpp.org/about/discuss.shtml> for a complete list.
Errata can be sent to <firstname.lastname@example.org>.
The following requirements keywords as used in this document are to be interpreted as described in RFC 2119: "MUST", "SHALL", "REQUIRED"; "MUST NOT", "SHALL NOT"; "SHOULD", "RECOMMENDED"; "SHOULD NOT", "NOT RECOMMENDED"; "MAY", "OPTIONAL".
10. The Internet Assigned Numbers Authority (IANA) is the central coordinator for the assignment of unique parameter values for Internet protocols, such as port numbers and URI schemes. For further information, see <http://www.iana.org/>.
11. The XMPP Registrar maintains a list of reserved protocol namespaces as well as registries of parameters used in the context of XMPP extension protocols approved by the XMPP Standards Foundation. For further information, see <https://xmpp.org/registrar/>.
Note: Older versions of this specification might be available at http://xmpp.org/extensions/attic/