Sublime Security low experimental mql

Link: .onion link with suspicious indicators

Detects inbound messages that reference onion/Tor addresses alongside extortion indicators such as cryptocurrency payment demands, language claiming stolen or exfiltrated data, Tor browser instructions, or a stolen filename used as the subject with no attachment. Also flags messages where an NLU classifier identifies malicious intent. Excludes solicited senders and highly trusted domains that pass DMARC authentication.

View Source

Detection Logic

type.inbound
and (
  any(body.links, .href_url.domain.tld == "onion")
  or regex.icontains(body.current_thread.text, '\b[a-z2-7]{56}\.onion\b')
)
and (
  sender.email.domain.root_domain in $free_email_providers
  // the actor speaks for themselves... "we hold copies", "we will publish it"
  or regex.icontains(strings.replace_confusables(body.current_thread.text),
                     '\bwe\s+(?:have
| hold
| obtained
| took
| stole
| downloaded
| exfiltrated
| copied
| (?:will
| are going to)\s+(?:publish
| release
| leak
| sell
| share
| expose
| send
| tell))\b'
  )
  or regex.icontains(body.current_thread.text,
                     '\btor\s*browser
| torproject\.org'
  )
  or regex.icontains(strings.replace_confusables(body.current_thread.text),
                     '\b(?:bitcoin
| btc
| monero
| xmr
| usdt)\b
| \bbc1q'
  )
  // a stolen file's name as the subject, with nothing attached
  or (
    regex.icontains(subject.subject, '\.(?:pdf
| xlsx?
| docx?
| csv
| zip)$')
    and length(attachments) == 0
  )
  or any(ml.nlu_classifier(strings.replace_confusables(body.current_thread.text)
         ).intents,
         .name != "benign" and .confidence != "low"
  )
)
and not profile.by_sender_email().solicited
// negate highly trusted sender domains unless they fail DMARC authentication
and not (
  sender.email.domain.root_domain in $high_trust_sender_root_domains
  and coalesce(headers.auth_summary.dmarc.pass, false)
)

Field Validations

Loading…

Comments (0)

Loading comments...