Detection rules › Splunk

Windows Credential Access From Browser Password Store

Author
Teoderick Contreras, Bhavin Patel Splunk
Source
upstream

The following analytic identifies a possible non-common browser process accessing its browser user data profile. This tactic/technique has been observed in various Trojan Stealers, such as SnakeKeylogger, which attempt to gather sensitive browser information and credentials as part of their exfiltration strategy. Detecting this anomaly can serve as a valuable pivot for identifying processes that access lists of browser user data profiles unexpectedly. This detection uses a lookup file browser_app_list that maintains a list of well known browser applications and the browser paths that are allowed to access the browser user data profiles.

MITRE ATT&CK coverage

TacticTechniques
DiscoveryT1012 Query Registry

Event coverage

ProviderEvent IDTitle
Security-Auditing4663An attempt was made to access an object.

Stages and Predicates

Stage 1: search

search EventCode=4663

Stage 2: stats

stats BY _time, object_file_path, object_file_name, dest, process_name, process_path, process_id, EventCode

Stage 3: lookup

lookup <lookup> browser_object_path, browser_process_name, isAllowed, object_file_path

Stage 4: stats

stats BY dest, process_name, process_path, process_id, EventCode, isAllowed

Stage 5: rex

rex field=process_name ...

Stage 6: eval

eval ... using (browser_process_name, extracted_process_name)

Stage 7: where

where isAllowed="false" isMalicious=1

Stage 8: search

search

Stage 9: search

search

Stage 10: search

search `macro`

Indicators

Each row is a field, operator, and value that the rule matches. The corpus column counts how many other rules in the catalog look for the same combination: high numbers point to widely-used, community-vetted indicators. Blank or 1 shows that the indicator is specific to this rule.

FieldKindValues
EventCodeeq
  • 4663 corpus 16 (splunk 16)
isAllowedeq
  • "false"
isMaliciouseq
  • 1

Neighbors

Stricter alternatives (narrower than this rule)

The rules below may be useful if you find the current rule is too noisy / lacks specificity.