Detection rules › Splunk

Splunk RCE via User XSLT

Status
production
Group by
action, clientip, decoded_field, host, uri, useragent
Author
Marissa Bower, Chase Franklin, Rod Soto, Bhavin Patel, Eric McGinnis, Splunk
Source
github.com/splunk/security_content

The following analytic identifies potential remote code execution (RCE) attempts via user-supplied Extensible Stylesheet Language Transformations (XSLT) in Splunk versions 9.1.x. It detects this activity by analyzing splunkd_ui logs for specific URI patterns and status codes indicative of XSLT injection attempts. This activity is significant because successful exploitation could allow an attacker to execute arbitrary code on the Splunk server. If confirmed malicious, this could lead to full system compromise, unauthorized data access, and further lateral movement within the network.

Known false positives

  • This search will provide information for investigation and hunting possible abuse of user-supplied XSLT. There may be false positives and results should individually evaluated. Please evaluate the source IP and useragent responsible for creating the requests.

MITRE ATT&CK coverage

TacticTechniques
Lateral Movement

Rule body

name: Splunk RCE via User XSLT
id: 6cb7e011-55fb-48e3-a98d-164fa854e37e
version: 7
creation_date: '2023-11-22'
modification_date: '2026-05-14'
author: Marissa Bower, Chase Franklin, Rod Soto, Bhavin Patel, Eric McGinnis, Splunk
status: production
type: Hunting
description: The following analytic identifies potential remote code execution (RCE) attempts via user-supplied Extensible Stylesheet Language Transformations (XSLT) in Splunk versions 9.1.x. It detects this activity by analyzing `splunkd_ui` logs for specific URI patterns and status codes indicative of XSLT injection attempts. This activity is significant because successful exploitation could allow an attacker to execute arbitrary code on the Splunk server. If confirmed malicious, this could lead to full system compromise, unauthorized data access, and further lateral movement within the network.
data_source: []
search: '`splunkd_ui` ((uri="*NO_BINARY_CHECK=1*" AND "*input.path=*.xsl*") OR uri="*dispatch*.xsl*") AND uri!= "*splunkd_ui*" | rex field=uri "(?<string>=\s*([\S\s]+))" | eval decoded_field=urldecode(string) | eval action=case(match(status,"200"),"Allowed",match(status,"303|500|401|403|404|301|406"),"Blocked",1=1,"Unknown") | stats count min(_time) as firstTime max(_time) as lastTime by clientip useragent uri decoded_field action host | rename clientip as src, uri as dest_uri | iplocation src | fillnull value="N/A" | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)` | table firstTime, lastTime src, useragent, action, count, Country, Region, City, dest_uri, decoded_field | `splunk_rce_via_user_xslt_filter`'
how_to_implement: This detection does not require you to ingest any new data. The detection does require the ability to search the _internal index.
known_false_positives: This search will provide information for investigation and hunting possible abuse of user-supplied XSLT. There may be false positives and results should individually evaluated. Please evaluate the source IP and useragent responsible for creating the requests.
references:
    - https://advisory.splunk.com/advisories/SVD-2023-1104
analytic_story:
    - Splunk Vulnerabilities
asset_type: Endpoint
cve:
    - CVE-2023-46214
mitre_attack_id:
    - T1210
product:
    - Splunk Enterprise
    - Splunk Enterprise Security
    - Splunk Cloud
category: application
security_domain: endpoint

Stages and Predicates

Stage 1: search

`splunkd_ui` ((uri="*NO_BINARY_CHECK=1*" AND "*input.path=*.xsl*") OR uri="*dispatch*.xsl*") AND uri!= "*splunkd_ui*"

Stage 2: rex

| rex field=uri "(?<string>=\s*([\S\s]+))"

Stage 3: eval

| eval decoded_field=urldecode(string)

Stage 4: eval

| eval action=case(match(status,"200"),"Allowed",match(status,"303|500|401|403|404|301|406"),"Blocked",1=1,"Unknown")
action =
ifmatch(status, "200")"Allowed"
elifmatch(status, "303|500|401|403|404|301|406")"Blocked"
else"Unknown"

Stage 5: stats

| stats count min(_time) as firstTime max(_time) as lastTime by clientip useragent uri decoded_field action host

Stage 6: rename

| rename clientip as src, uri as dest_uri

Stage 7: search

| iplocation src

Stage 8: fillnull

| fillnull value="N/A"

Stage 9: search

| `security_content_ctime(firstTime)`

Stage 10: search

| `security_content_ctime(lastTime)`

Stage 11: table

| table firstTime, lastTime src, useragent, action, count, Country, Region, City, dest_uri, decoded_field

Stage 12: search

| `splunk_rce_via_user_xslt_filter`

Indicators

These rows show field, operator, and value matches.

FieldKindValuesSearch
sourcetypeeq
  • splunkd_ui_access
field:"sourcetype" kind:eq value:"splunkd_ui_access"
urieq
  • "*NO_BINARY_CHECK=1*"
  • "*dispatch*.xsl*"
field:"uri" kind:eq
urine
  • "*splunkd_ui*"
field:"uri" kind:ne

Search terms

These SPL tokens match against raw event text.

StageTerm
1"*input.path=*.xsl*"
7iplocation
7src