When Credential Harvesting/Dumping is observed, analysts must determine which users were logged in at the time and might have had their credentials compromised.

The Splunk SIEM app includes a Run Live Query Alert Action.

The first search runs as an Alert in Splunk SIEM on a regular basis, perhaps every 5 – 10 minutes, and identifies any possible instances of credential theft. It also defines the Live Query SQL to run.

eventtype="vmware_cbc_cb_analytics" threat_indicators{}.ttps{} = "MITRE_T1003_OS_CREDENTIAL_DUMP"
| eval query_logged_in_users = "SELECT * FROM logged_in_users;"
| dedup device_id

Each result triggers the VMware CBC Run Livequery Alert Action using the following parameters:

Splunk SIEM Run Livequery window

When Carbon Black Cloud detects credential scraping, Splunk SIEM will automatically figure out who was logged in to the impacted endpoints, and bring in those results through the App's built-in Live Query input. You can then run a Splunk query to combine the original alerts with the logged in user information.

eventtype="vmware_cbc_cb_analytics" threat_indicators{}.ttps{} = "MITRE_T1003_OS_CREDENTIAL_DUMP"
| dedup id
| reltime
| join device_id type=outer [
  search eventtype="vmware_cbc_base_index" sourcetype="vmware:cbc:livequery:result" query_name="Splunk: Logged In Users (memory scraping)" status=matched
  | rename device.* as device_*, fields.* as *
  | eval _time = time
  | reltime
  | strcat "User=" user ", type=" tty ", SID=" sid ", " reltime login_string
  | stats values(login_string) as logins by device_id
]
| table device_id, device_name, reltime, sensor_action, threat_cause_actor_name, reason, logins

Query result of combined alerts with logged in users