Detection rules › Splunk
Kubernetes Suspicious Image Pulling
The following analytic detects suspicious image pulling in Kubernetes environments. It identifies this activity by monitoring Kubernetes audit logs for image pull requests that do not match a predefined list of allowed images. This behavior is significant for a SOC as it may indicate an attacker attempting to deploy malicious software or infiltrate the system. If confirmed malicious, the impact could be severe, potentially leading to unauthorized access to sensitive systems or data, and enabling further malicious activities within the cluster.
MITRE ATT&CK coverage
| Tactic | Techniques |
|---|---|
| Discovery | T1526 Cloud Service Discovery |
Rule body splunk
name: Kubernetes Suspicious Image Pulling
id: 4d3a17b3-0a6d-4ae0-9421-46623a69c122
version: 10
creation_date: '2023-12-20'
modification_date: '2026-05-13'
author: Patrick Bareiss, Splunk
status: production
type: Anomaly
description: The following analytic detects suspicious image pulling in Kubernetes environments. It identifies this activity by monitoring Kubernetes audit logs for image pull requests that do not match a predefined list of allowed images. This behavior is significant for a SOC as it may indicate an attacker attempting to deploy malicious software or infiltrate the system. If confirmed malicious, the impact could be severe, potentially leading to unauthorized access to sensitive systems or data, and enabling further malicious activities within the cluster.
data_source:
- Kubernetes Audit
search: |-
`kube_audit` requestObject.message="Pulling image*"
| search NOT `kube_allowed_images`
| fillnull
| stats count
BY objectRef.name objectRef.namespace objectRef.resource
requestReceivedTimestamp requestURI responseStatus.code
sourceIPs{} stage user.groups{}
user.uid user.username userAgent
verb
| rename sourceIPs{} as src_ip, user.username as user
| `kubernetes_suspicious_image_pulling_filter`
how_to_implement: The detection is based on data that originates from Kubernetes Audit logs. Ensure that audit logging is enabled in your Kubernetes cluster. Kubernetes audit logs provide a record of the requests made to the Kubernetes API server, which is crucial for monitoring and detecting suspicious activities. Configure the audit policy in Kubernetes to determine what kind of activities are logged. This is done by creating an Audit Policy and providing it to the API server. Use the Splunk OpenTelemetry Collector for Kubernetes to collect the logs. This doc will describe how to collect the audit log file https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/migration-from-sck.md. When you want to use this detection with AWS EKS, you need to enable EKS control plane logging https://docs.aws.amazon.com/eks/latest/userguide/control-plane-logs.html. Then you can collect the logs from Cloudwatch using the AWS TA https://splunk.github.io/splunk-add-on-for-amazon-web-services/CloudWatchLogs/.
known_false_positives: No false positives have been identified at this time.
references:
- https://kubernetes.io/docs/tasks/debug/debug-cluster/audit/
drilldown_searches:
- name: View the detection results for - "$user$"
search: '%original_detection_search% | search user = "$user$"'
earliest_offset: $info_min_time$
latest_offset: $info_max_time$
- name: View risk events for the last 7 days for - "$user$"
search: '| from datamodel Risk.All_Risk | search normalized_risk_object IN ("$user$") | stats count min(_time) as firstTime max(_time) as lastTime values(search_name) as "Search Name" values(risk_message) as "Risk Message" values(analyticstories) as "Analytic Stories" values(annotations._all) as "Annotations" values(annotations.mitre_attack.mitre_tactic) as "ATT&CK Tactics" by normalized_risk_object | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)`'
earliest_offset: 7d
latest_offset: "0"
intermediate_findings:
entities:
- field: user
type: user
score: 20
message: Suspicious image $objectRef.name$ pulled in Kubernetes from ip $src_ip$ by user $user$
threat_objects:
- field: src_ip
type: ip_address
analytic_story:
- Kubernetes Security
asset_type: Kubernetes
mitre_attack_id:
- T1526
product:
- Splunk Enterprise
- Splunk Enterprise Security
- Splunk Cloud
category: cloud
security_domain: network
tests:
- name: True Positive Test
attack_data:
- data: https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1526/kubernetes_audit_pull_image/kubernetes_audit_pull_image.json
sourcetype: _json
source: kubernetes
test_type: unit
Stages and Predicates
Stage 1: search
`kube_audit` requestObject.message="Pulling image*"
Stage 2: search
| search NOT `kube_allowed_images`
Stage 3: fillnull
| fillnull
Stage 4: stats
| stats count
BY objectRef.name objectRef.namespace objectRef.resource
requestReceivedTimestamp requestURI responseStatus.code
sourceIPs{} stage user.groups{}
user.uid user.username userAgent
verb
Stage 5: rename
| rename sourceIPs{} as src_ip, user.username as user
Stage 6: search
| `kubernetes_suspicious_image_pulling_filter`
Exclusions
Top-level NOT(...) conjuncts: predicates this rule actively suppresses.
| Field | Kind | Excluded values |
|---|---|---|
objectRef.name | in | "*falco*", "*splunk*" |
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.
| Field | Kind | Values |
|---|---|---|
requestObject.message | eq |
|