Brand impersonation: SharePoint PDF attachment with credential theft language

name: "Brand impersonation: SharePoint PDF attachment with credential theft language"
description: "PDF attachment contains SharePoint logo and high-confidence credential theft language detected via OCR analysis. The attachment includes URLs and originates from an unsolicited or low-reputation sender, excluding legitimate SharePoint file sharing notifications."
type: "rule"
severity: "medium"
source: |
  type.inbound
  and (
    any(attachments,
        (
          .file_type == "pdf"
          and any(ml.logo_detect(.).brands, .name == "Microsoft SharePoint")
          and any(file.explode(.), length(.scan.url.urls) > 0)
          and any(file.explode(.),
                  any(ml.nlu_classifier(.scan.ocr.raw).intents,
                      .name == "cred_theft" and .confidence == "high"
                  )
          )
        )
    )
  )
  // negate sharepoint file share
  and not (
    // based on the message id format
    (
      (
        strings.starts_with(headers.message_id, '<Share-')
        and strings.ends_with(headers.message_id, '@odspnotify>')
      )
      or // negate legitimate access request to file
   (
        strings.starts_with(headers.message_id, '<Sharing')
        and strings.ends_with(headers.message_id, '@odspnotify>')
      )
      // deal with Google thinking the message ID is "broke"
      or (
        strings.icontains(headers.message_id, 'SMTPIN_ADDED_BROKEN')
        and any(headers.hops,
                any(.fields,
                    .name == "X-Google-Original-Message-ID"
                    and strings.starts_with(.value, '<Share-')
                    and strings.ends_with(.value, '@odspnotify>')
                )
        )
      )
    )
    // all of the "action" links are sharepoint/ms
    and all(filter(body.links,
                   strings.icontains(subject.subject, .display_text)
                   or .display_text == "Open"
            ),
            .href_url.domain.root_domain in ("sharepoint.com")
            or (
              .href_url.domain.tld == "ms"
              // Microsoft does not own the .ms TLD, this checks to ensure it is one of their domains
              and (
                network.whois(.href_url.domain).registrant_company == "Microsoft Corporation"
                or strings.ilike(network.whois(.href_url.domain).registrar_name,
                                 "*MarkMonitor*",
                                 "*CSC Corporate*",
                                 "*com laude*"
                )
              )
            )
    )
  )
  and not (
    (
      (subject.is_reply or subject.is_forward)
      and (
        (length(headers.references) > 0 or headers.in_reply_to is not null)
        // ensure that there are actual threads
        and (
          length(body.previous_threads) > 0
          or (length(body.html.display_text) - length(body.current_thread.text)) > 200
        )
      )
    )
  )
  and (
    profile.by_sender_email().prevalence != 'common'
    or not profile.by_sender_email().solicited
    or profile.by_sender().any_messages_malicious_or_spam
  )
  and not profile.by_sender().any_messages_benign
  
  // negate highly trusted sender domains unless they fail DMARC authentication
  and (
    (
      sender.email.domain.root_domain in $high_trust_sender_root_domains
      and not headers.auth_summary.dmarc.pass
    )
    or sender.email.domain.root_domain not in $high_trust_sender_root_domains
  )

attack_types:
  - "Credential Phishing"
tactics_and_techniques:
  - "Impersonation: Brand"
  - "Social engineering"
  - "PDF"
  - "Evasion"
detection_methods:
  - "Computer Vision"
  - "File analysis"
  - "Natural Language Understanding"
  - "Optical Character Recognition"
  - "Sender analysis"
  - "URL analysis"
  - "Header analysis"
  - "Whois"
id: "ae3756fa-3751-5fba-b68d-d91164fd359c"