Fake thread with suspicious indicators

  1name: "Fake thread with suspicious indicators"
  2description: "Fake thread contains suspicious indicators, which can lead to BEC, credential phishing, and other undesirable outcomes."
  3type: "rule"
  4severity: "medium"
  5source: |
  6  type.inbound
  7  // fake thread check
  8  and (length(headers.references) == 0 or headers.in_reply_to is null)
  9  and (
 10    subject.is_reply
 11    or subject.is_forward
 12    // fake thread, but no indication in the subject line
 13    // current_thread pulls the recent thread, but the full body contains the fake "original" email
 14    or (
 15      not (subject.is_reply or subject.is_forward)
 16      and any([body.current_thread.text, body.html.display_text, body.plain.raw],
 17              3 of (
 18                strings.icontains(., "from:"),
 19                strings.icontains(., "to:"),
 20                strings.icontains(., "sent:"),
 21                strings.icontains(., "date:"),
 22                strings.icontains(., "cc:"),
 23                strings.icontains(., "subject:")
 24              )
 25      )
 26      and length(body.current_thread.text) + 100 < length(coalesce(body.html.display_text,
 27                                                                   body.plain.raw
 28                                                          )
 29      )
 30    )
 31  )
 32  
 33  // negating bouncebacks
 34  and not any(attachments,
 35              .content_type in ("message/delivery-status", "message/rfc822")
 36  )
 37  // negating Google Calendar invites
 38  and (
 39    (
 40      headers.return_path.domain.domain is not null
 41      and headers.return_path.domain.domain != 'calendar-server.bounces.google.com'
 42    )
 43    or headers.return_path.domain.domain is null
 44  )
 45  // not mimecast secure message from internal source
 46  and not (
 47    strings.istarts_with(headers.message_id, '<Mimecast.')
 48    and strings.iends_with(headers.message_id, '.mimecast.lan>')
 49    and headers.hops[0].received.server.raw == "relay.mimecast.com"
 50    and strings.icontains(headers.hops[0].received.source.raw, 'mimecast.lan')
 51  )
 52  
 53  // and not solicited
 54  and not profile.by_sender().solicited
 55  and 4 of (
 56    // language attempting to engage
 57    (
 58      any(ml.nlu_classifier(body.current_thread.text).entities,
 59          .name == "request"
 60      )
 61      and any(ml.nlu_classifier(body.current_thread.text).entities,
 62              .name == "financial"
 63      )
 64    ),
 65  
 66    // invoicing language
 67    (
 68      any(ml.nlu_classifier(body.current_thread.text).tags, .name == "invoice")
 69      or any(ml.nlu_classifier(body.current_thread.text).entities,
 70             .text == "invoice"
 71      )
 72    ),
 73  
 74    // urgency request
 75    any(ml.nlu_classifier(body.current_thread.text).entities, .name == "urgency"),
 76  
 77    // cred_theft detection
 78    any(ml.nlu_classifier(body.current_thread.text).intents,
 79        .name == "cred_theft" and .confidence in~ ("medium", "high")
 80    ),
 81  
 82    // commonly abused sender TLD
 83    strings.ilike(sender.email.domain.tld, "*.jp"),
 84  
 85    // headers traverse abused TLD
 86    any(headers.domains, strings.ilike(.tld, "*.jp")),
 87  
 88    // known suspicious pattern in the URL path
 89    any(body.links, regex.match(.href_url.path, '\/[a-z]{3}\d[a-z]')),
 90  
 91    // link display text is in all caps
 92    any(body.links, regex.match(.display_text, '[A-Z ]+')),
 93  
 94    // link display text contains invisible characters (U+200F)
 95    any(body.links, strings.contains(.display_text, "\u{200F}")),
 96  
 97    // Low reputation link with display text ending in a document extension
 98    any(body.links,
 99        .href_url.domain.root_domain not in $tranco_1m
100        and .href_url.domain.valid
101        and .href_url.domain.root_domain not in $org_domains
102        and .href_url.domain.root_domain not in $high_trust_sender_root_domains
103        and (
104          any($file_extensions_macros, strings.ends_with(..display_text, .))
105          or strings.ends_with(.display_text, 'pdf')
106        )
107    ),
108  
109    // display name contains an email
110    regex.contains(sender.display_name, '[a-z0-9]+@[a-z]+'),
111  
112    // Sender domain is empty
113    sender.email.domain.domain == "",
114  
115    // sender domain matches no body domains
116    all(body.links,
117        .href_url.domain.root_domain != sender.email.domain.root_domain
118    ),
119  
120    // body contains name of VIP
121    (
122      any($org_vips, strings.icontains(body.html.inner_text, .display_name))
123      or any($org_vips, strings.icontains(body.plain.raw, .display_name))
124    ),
125  
126    // new body domain
127    any(body.links, network.whois(.href_url.domain).days_old < 30),
128  
129    // new sender domain
130    network.whois(sender.email.domain).days_old < 30,
131  
132    // new sender
133    profile.by_sender().days_known < 7,
134  
135    // excessive whitespace
136    (
137      regex.icontains(body.html.raw, '((<br\s*/?>\s*){20,}|\n{20,})')
138      or regex.icontains(body.html.raw, '(<p[^>]*>\s*<br\s*/?>\s*</p>\s*){30,}')
139      or regex.icontains(body.html.raw,
140                         '(<p class=".*?"><span style=".*?"><o:p>&nbsp;</o:p></span></p>\s*){30,}'
141      )
142      or regex.icontains(body.html.raw, '(<p>&nbsp;</p>\s*){7,}')
143      or regex.icontains(body.html.raw, '(<p>&nbsp;</p><br>\s*){7,}')
144      or regex.icontains(body.html.raw, '(<p[^>]*>\s*&nbsp;<br>\s*</p>\s*){5,}')
145      or regex.icontains(body.html.raw, '(<p[^>]*>&nbsp;</p>\s*){7,}')
146    ),
147  
148    // body contains recipient SLD
149    any(recipients.to,
150        strings.icontains(body.current_thread.text, .email.domain.sld)
151    )
152  )
153  
154  // negate highly trusted sender domains unless they fail DMARC authentication
155  and (
156    (
157      sender.email.domain.root_domain in $high_trust_sender_root_domains
158      and not headers.auth_summary.dmarc.pass
159    )
160    or sender.email.domain.root_domain not in $high_trust_sender_root_domains
161  )
162  and not profile.by_sender().any_messages_benign  
163tags:
164  - "Attack surface reduction"
165attack_types:
166  - "BEC/Fraud"
167  - "Credential Phishing"
168  - "Spam"
169tactics_and_techniques:
170  - "Evasion"
171  - "Social engineering"
172detection_methods:
173  - "Content analysis"
174  - "Header analysis"
175  - "Natural Language Understanding"
176  - "Sender analysis"
177id: "c2e18a57-1f52-544f-bb6d-a578e286cf89"