• Sublime Core Feed
High Severity

Link: Adobe Share with Suspicious Indicators

Labels

Credential Phishing
Evasion
Free file host
Content analysis
URL screenshot
Sender analysis
Natural Language Understanding
URL analysis

Description

The detection rule matches messages sent from Adobe and contain indicators of malicious use. The indicators include observed call to action phrases, suspicious filenames, all capital filenames, the sender's display name (as determined by NLU) included in the comment section, or Microsoft branding on the shared link.

References

No references.

Sublime Security
Created Nov 13th, 2024 • Last updated Dec 3rd, 2024
Feed Source
Sublime Core Feed
Source
GitHub
type.inbound
// from Adobe Actual
and strings.icontains(sender.display_name, 'via Adobe')
and sender.email.email == 'message@adobe.com'
and headers.auth_summary.dmarc.pass
// contains a link to open or review a share
and any(body.links, .display_text =~ "open" or .display_text =~ "review")

// not sent from a Adobe User within the org's domains
and not any($org_domains,
            strings.icontains(sender.display_name,
                              strings.concat("@", ., ' via Adobe')
            )
            // sometimes the email is in parentheses
            or strings.icontains(sender.display_name,
                              strings.concat("@", ., ') via Adobe')
            )
)
and (
  // the comments observed wording, using the html to make sure it's in the actor controlled section of the message
  regex.icontains(body.html.raw,
                  '<tr>[\r\n]+<td style="color:#505050; font-family:adobe-clean, Helvetica Neue, Helvetica, Verdana, Arial, sans-serif; font-size:18px; line-height:26px; padding-top:20px;">[\r\n]+<xmp style="font-family:adobe-clean, Helvetica Neue, Helvetica, Verdana, Arial, sans-serif; font-size:18px; line-height:26px overflow-x:auto; white-space:pre-wrap; white-space:-moz-pre-wrap; white-space:-pre-wrap; white-space:-o-pre-wrap; word-wrap:break-word;">Please review the attached below for your reference,'
  )
  // the filename shared
  or regex.icontains(body.html.raw,
                     // , ends in some random numbers
                     '<td style="color:#000000; font-family:adobe-clean, Helvetica Neue, Helvetica, Verdana, Arial, sans-serif; font-size:24px; line-height:26px; padding-top:65px;">[\r\n]+<strong>[^\<]+<\/strong> (invited you to review|has shared) <strong>[^\<]+([]|[[:punct:]\s](?:AP|AR)?\d+[a-z]?)<\/strong></td>[\r\n]+</tr>'
  )
  // contains all capital letters, allowing for numbers
  or regex.contains(body.html.raw,
                    '<td style="color:#000000; font-family:adobe-clean, Helvetica Neue, Helvetica, Verdana, Arial, sans-serif; font-size:24px; line-height:26px; padding-top:65px;">[\r\n]+<strong>[^\<]+<\/strong> (invited you to review|has shared) <strong>(?:[A-Z0-9_\-\s]+)<\/strong></td>[\r\n]+</tr>',
  )
  // contains commonly observed themes used by actors
  or regex.icontains(body.html.raw,
                     // , ends in some random numbers
                     '<td style=\"color:#000000; font-family:adobe-clean, Helvetica Neue, Helvetica, Verdana, Arial, sans-serif; font-size:24px; line-height:26px; padding-top:65px;\">[\r\n]+<strong>[^\<]+<\/strong> invited you to review <strong>[^\<]*(Invoice|Payment|Agreement|Settlements|Overdue|Confidential|Transaction)[^\<]*<\/strong></td>[\r\n]+</tr>')
  // the NLU detected "sender" is included within the body wrapped with new lines indicating it's a "signature"
  or any(filter(ml.nlu_classifier(body.current_thread.text).entities,
                .name == "sender" and .text not in ('Customer Support', 'SHARED ON')
                // in some cases the filename is detected as the sender
                // we can filter out this case when the detected "sender"
                // text is the file shared
                and not strings.icontains(body.current_thread.text,
                                          strings.concat("invited you to edit\n",
                                                         .text,
                                                         "\nOpen"
                                          )
                )
         ),
         strings.icontains(body.current_thread.text,
                           strings.concat("\n", .text, "\n")
         )
  )
  // finally we'll hit the actual page and see if we can get some enrichment functions to give up some gold
  or any(filter(body.links, .display_text =~ "open" or .display_text =~ "review"),
    // detected as Microsoft
    any(ml.logo_detect(ml.link_analysis(., mode="aggressive").screenshot).brands, .name in ("Microsoft") and .confidence == "high")
  )
)
MQL Rule Console
DocsLearning Labs

Playground

Test against your own EMLs or sample data.

Share

Post about this on your socials.

Get Started. Today.

Managed or self-managed. No MX changes.

Get Started