Extortion / sextortion (untrusted sender)
Detects extortion and sextortion attempts by analyzing the email body text from an untrusted sender.
Sublime rule (View on GitHub)
  1name: "Extortion / sextortion (untrusted sender)"
  2description: |
  3    Detects extortion and sextortion attempts by analyzing the email body text from an untrusted sender.
  4references:
  5  - "https://krebsonsecurity.com/2018/07/sextortion-scam-uses-recipients-hacked-passwords/"
  6type: "rule"
  7severity: "low"
  8source: |
  9  type.inbound
 10  and length(filter(body.links, .display_text is not null)) < 10
 11  and not (
 12    ml.nlu_classifier(body.current_thread.text).language == "english"
 13    and any(ml.nlu_classifier(body.html.display_text).topics,
 14            .name in (
 15              "News and Current Events",
 16              "Newsletters and Digests",
 17              "Advertising and Promotions"
 18            )
 19            and .confidence in ("high", "medium")
 20    )
 21  )
 22  and (
 23    (
 24      any(ml.nlu_classifier(strings.replace_confusables(body.current_thread.text)).intents,
 25          (.name == "extortion" and .confidence == "high")
 26      )
 27      and (
 28        any(ml.nlu_classifier(strings.replace_confusables(body.current_thread.text
 29                              )
 30            ).entities,
 31            .name == "financial"
 32            or (
 33              .name is not null
 34              and regex.icontains(.text,
 35                                  "cybḛ[rŗřṙȑȓɍʀɼɽг]c[rŗřṙȑȓɍʀɼɽг]imina[lĺļľḷḹḽłƖʟḻ]s"
 36              )
 37            )
 38        )
 39        or any(ml.nlu_classifier(strings.replace_confusables(body.current_thread.text
 40                                 )
 41               ).topics,
 42               .name == "Financial Communications" and .confidence != "low"
 43        )
 44      )
 45    )
 46    // manual indicators failsafe
 47    or 3 of (
 48      // malware terms
 49      regex.icontains(strings.replace_confusables(body.current_thread.text),
 50                      "(?:(?:spy|[mṁ][aȁ]l)[wŵ][aȁ][rŗ]e|[tŢ][rŗȓ][oố]j[aǻ][nņ]|[rȓ]emote (?:entry|cont[rř]o[lĺ])|infiltrat(?:ed|ion)|backdoor|vi[rṙ]us|intruder|(?:your|the).{0,15}(?:device|system|computer|phone).{0,10}(?:became|was|got|is).{0,5}comprom[ḯiïíįī]sed|prov[ḯiïíįī]d[ḯiïíįī]ng.{0,20}full [aảǡą]ccess)"
 51      ),
 52      // actions recorded
 53      regex.icontains(strings.replace_confusables(body.current_thread.text),
 54                      "(?:p[oộ][rŗ]n|a[dȡ]ult (?:web)?site|webcam|mastu[rŗ]bating|je[rŗ]king off|pleasu[rŗȑ]i[nŋ]g you[rŗṛ]self|getting off|expli[cƈ]it|cl[ḯiïíįī]ps.{0,20}screenshots)"
 55      ),
 56      regex.icontains(strings.replace_confusables(body.current_thread.text),
 57                      "(?:pe[rŗ]ve[rŗ]t|pe[rŗ]ve[rŗ]sion|mastu[rŗ]bat)"
 58      ),
 59      // a timeframe to pay
 60      regex.icontains(strings.replace_confusables(body.current_thread.text),
 61                      '[ilo0-9]{2} (?:hou[rŗṝ][sṣ]|uu[rŗ])',
 62                      '(?:one|tw[oờ]|2|th[rŗ]ee|\d) [dḍ][aảǡą]y[sṣ]?'
 63      ),
 64      // a promise from the actor
 65      regex.icontains(strings.replace_confusables(body.current_thread.text),
 66                      '(?:pe[rŗ]manently|will|I''ll) delete|([rŗ]emove|destroy) (?:[\p{L}\p{M}\p{N}]+\s*){0,4} (?:data|ev[ḯiïįīí]dence|v[ḯiïíįī]deos?)'
 67      ),
 68      // a threat from the actor
 69      regex.icontains(strings.replace_confusables(body.current_thread.text),
 70                      '(?:\bsen[dt]|forward|expose)\s*(?:[\p{L}\p{N}]+\s*){0,5}\s*to\s*(?:[\p{L}\p{N}]+\s*){0,5}(?:contacts|media|family|f[rŗ]iends|coworkers|associates)'
 71      ),
 72      // bitcoin language (excluding newsletters)
 73      (
 74        regex.icontains(strings.replace_confusables(body.current_thread.text),
 75                        '[bḆḂ]it[cĉƈć][oöة]+in|\bbtc\b|blockchain'
 76        )
 77        // negate cryptocurrency newsletters
 78        and not (
 79          any(body.links,
 80              strings.icontains(.display_text, "unsubscribe")
 81              and (
 82                strings.icontains(.href_url.path, "unsubscribe")
 83                // handle mimecast URL rewrites
 84                or (
 85                  .href_url.domain.root_domain == 'mimecastprotect.com'
 86                  and strings.icontains(.href_url.query_params,
 87                                        sender.email.domain.root_domain
 88                  )
 89                )
 90              )
 91          )
 92        )
 93      ),
 94      (
 95        regex.icontains(strings.replace_confusables(body.current_thread.text),
 96                        '(?:contact the police|(?:bitcoin|\bbtc\b).{0,20}wallet)'
 97        )
 98        and regex.icontains(strings.replace_confusables(body.current_thread.text),
 99                            '(?:\b[13][a-km-zA-HJ-NP-Z0-9]{24,34}\b)|\bX[1-9A-HJ-NP-Za-km-z]{33}\b|\b(?:0x[a-fA-F0-9]{40})\b|\b[LM3][a-km-zA-HJ-NP-Z1-9]{26,33}\b|\b[48][0-9AB][1-9A-HJ-NP-Za-km-z]{93}\b'
100        )
101      ),
102      regex.icontains(strings.replace_confusables(body.current_thread.text),
103                      'bc1q.{0,50}\b'
104      )
105    )
106  )
107  and (
108    not profile.by_sender().solicited
109    or (
110      profile.by_sender().any_messages_malicious_or_spam
111      and not profile.by_sender().any_messages_benign
112    )
113    or any(headers.hops, any(.fields, .name == "X-Google-Group-Id"))
114  
115    // many extortion emails spoof sender domains and fail sender authentication
116    or (
117      not headers.auth_summary.dmarc.pass
118      or headers.auth_summary.dmarc.pass is null
119      or not headers.auth_summary.spf.pass
120    )
121  )
122  
123  // negate benign newsletters that mention cyber extortion
124  and not (
125    any(body.links,
126        strings.icontains(.display_text, "unsubscribe")
127        and strings.icontains(.href_url.path, "unsubscribe")
128    )
129    // newsletters are typically longer than the average extortion script
130    and length(body.current_thread.text) > 2000
131  )
132  and length(body.current_thread.text) < 8000
133  // negate highly trusted sender domains unless they fail DMARC authentication
134  and (
135    (
136      sender.email.domain.root_domain in $high_trust_sender_root_domains
137      and not headers.auth_summary.dmarc.pass
138    )
139    or sender.email.domain.root_domain not in $high_trust_sender_root_domains
140  )  
141
142attack_types:
143  - "Extortion"
144tactics_and_techniques:
145  - "Social engineering"
146  - "Spoofing"
147detection_methods:
148  - "Content analysis"
149  - "Header analysis"
150  - "Natural Language Understanding"
151  - "Sender analysis"
152id: "265913eb-2ccd-5f77-9a09-f6d8539fd2f6"