Kubernetes Data Copy via kubectl cp

Severity: medium
Group by: name, namespace, username
Log types: Amazon.EKS.Audit, Azure.MonitorActivity, GCP.AuditLog
Tags: Kubernetes, Exfiltration, Data Theft, Credential Access, Unified Detection
Reference: https://kubernetes.io/docs/reference/generated/kubectl/kubectl-commands#cp
Source: github.com/panther-labs/panther-analysis

This detection monitors for kubectl cp operations that copy files from pods to local machines, which can indicate data exfiltration. When kubectl cp is used to copy files from a pod, it executes a tar command with stdout output (tar cf -) inside the container and streams the data back through the Kubernetes API server. Attackers who gain cluster access can use this technique to steal application secrets, credentials, configuration files, or sensitive data from container filesystems without leaving obvious traces inside the pod itself. While kubectl cp has legitimate uses for debugging and backup, unexpected usage should be investigated.

MITRE ATT&CK coverage

Tactic	Techniques
Credential Access	`T1552` Unsecured Credentials
Collection	`T1530` Data from Cloud Storage

Rule body yaml

AnalysisType: rule
RuleID: "Kubernetes.Kubectl.CP.Operation"
DisplayName: "Kubernetes Data Copy via kubectl cp"
Enabled: true
Filename: k8s_kubectl_cp_operation.py
LogTypes:
  - Amazon.EKS.Audit
  - Azure.MonitorActivity
  - GCP.AuditLog
Tags:
  - Kubernetes
  - Exfiltration
  - Data Theft
  - Credential Access
  - Unified Detection
Severity: Medium
Description: >
  This detection monitors for kubectl cp operations that copy files from pods to local machines,
  which can indicate data exfiltration. When kubectl cp is used to copy files from a pod, it
  executes a tar command with stdout output (tar cf -) inside the container and streams the data
  back through the Kubernetes API server. Attackers who gain cluster access can use this technique
  to steal application secrets, credentials, configuration files, or sensitive data from container
  filesystems without leaving obvious traces inside the pod itself. While kubectl cp has legitimate
  uses for debugging and backup, unexpected usage should be investigated.
Runbook: |
  1. Review all exec and kubectl cp operations by this user in the 24 hours before and after the alert
  2. Identify what files or directories were exfiltrated and assess their sensitivity
  3. Search for other suspicious API activity from this user or service account across all clusters in the past 7 days
Reports:
  MITRE ATT&CK:
    - TA0010:T1530 # Exfiltration: Data from Cloud Storage
    - TA0006:T1552 # Credential Access: Unsecured Credentials
Reference: https://kubernetes.io/docs/reference/generated/kubectl/kubectl-commands#cp
DedupPeriodMinutes: 60
SummaryAttributes:
  - username
  - namespace
  - name
  - p_source_label
Tests:
  - Name: EKS kubectl cp from pod (requestURI format)
    ExpectedResult: true
    Log:
      {
        "kind": "Event",
        "apiVersion": "audit.k8s.io/v1",
        "verb": "create",
        "user": {"username": "attacker@example.com"},
        "sourceIPs": ["203.0.113.42"],
        "userAgent": "kubectl/v1.28.0",
        "objectRef": {
          "resource": "pods",
          "subresource": "exec",
          "namespace": "production",
          "name": "webapp-pod",
          "apiVersion": "v1"
        },
        "responseStatus": {"code": 200},
        "requestURI": "/api/v1/namespaces/production/pods/webapp-pod/exec?command=tar&command=cf&command=-&command=/app/secrets/credentials.json&container=webapp&stdout=true",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit",
        "p_source_label": "eks-cluster"
      }
  - Name: AKS kubectl cp from pod with SSH keys
    ExpectedResult: true
    Log:
      {
        "p_log_type": "Azure.MonitorActivity",
        "category": "kube-audit",
        "operationName": "Microsoft.ContainerService/managedClusters/diagnosticLogs/Read",
        "properties": {
          "log": "{\"kind\":\"Event\",\"apiVersion\":\"audit.k8s.io/v1\",\"verb\":\"create\",\"user\":{\"username\":\"malicious-user@example.com\"},\"sourceIPs\":[\"1.2.3.4\"],\"objectRef\":{\"resource\":\"pods\",\"subresource\":\"exec\",\"namespace\":\"default\",\"name\":\"api-server\"},\"responseStatus\":{\"code\":200},\"requestObject\":{\"command\":[\"tar\",\"cf\",\"-\",\"/root/.ssh/id_rsa\"],\"container\":\"api\"}}"
        },
        "p_source_label": "aks-cluster"
      }
  - Name: GKE kubectl cp from pod
    ExpectedResult: true
    Log:
      {
        "protoPayload": {
          "authenticationInfo": {"principalEmail": "user@company.com"},
          "authorizationInfo": [{
            "granted": true,
            "permission": "io.k8s.core.v1.pods.exec",
            "resource": "core/v1/namespaces/production/pods/database/exec"
          }],
          "methodName": "io.k8s.core.v1.pods.exec.create",
          "requestMetadata": {"callerIP": "8.8.8.8"},
          "resourceName": "core/v1/namespaces/production/pods/database/exec",
          "serviceName": "k8s.io",
          "request": {
            "command": ["tar", "cf", "-", "/var/lib/postgresql/data/config"],
            "container": "postgres"
          }
        },
        "resource": {
          "type": "k8s_cluster",
          "labels": {"project_id": "test-project"}
        },
        "p_log_type": "GCP.AuditLog",
        "p_source_label": "gke-cluster"
      }
  - Name: EKS kubectl cp TO pod (not exfiltration)
    ExpectedResult: false
    Log:
      {
        "kind": "Event",
        "verb": "create",
        "objectRef": {"resource": "pods", "subresource": "exec", "namespace": "default", "name": "app-pod"},
        "responseStatus": {"code": 200},
        "requestURI": "/api/v1/namespaces/default/pods/app-pod/exec?command=tar&command=xf&command=-&command=-C&command=/tmp&container=app",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit"
      }
  - Name: EKS regular tar archive creation
    ExpectedResult: false
    Log:
      {
        "kind": "Event",
        "verb": "create",
        "objectRef": {"resource": "pods", "subresource": "exec", "namespace": "default", "name": "backup-pod"},
        "responseStatus": {"code": 200},
        "requestURI": "/api/v1/namespaces/default/pods/backup-pod/exec?command=tar&command=czf&command=backup.tar.gz&command=/app/data",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit"
      }
  - Name: EKS regular exec without tar
    ExpectedResult: false
    Log:
      {
        "kind": "Event",
        "verb": "create",
        "user": {"username": "admin@example.com"},
        "objectRef": {"resource": "pods", "subresource": "exec", "namespace": "default", "name": "debug-pod"},
        "responseStatus": {"code": 200},
        "requestURI": "/api/v1/namespaces/default/pods/debug-pod/exec?command=/bin/sh&stdin=true&tty=true",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit"
      }
  - Name: EKS system namespace (excluded)
    ExpectedResult: false
    Log:
      {
        "kind": "Event",
        "verb": "create",
        "user": {"username": "system:serviceaccount:kube-system:debug"},
        "objectRef": {"resource": "pods", "subresource": "exec", "namespace": "kube-system", "name": "debug-pod"},
        "responseStatus": {"code": 200},
        "requestURI": "/api/v1/namespaces/kube-system/pods/debug-pod/exec?command=tar&command=cf&command=-&command=/var/log",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit"
      }
  - Name: EKS failed request (excluded)
    ExpectedResult: false
    Log:
      {
        "kind": "Event",
        "verb": "create",
        "user": {"username": "attacker@example.com"},
        "objectRef": {"resource": "pods", "subresource": "exec", "namespace": "default", "name": "secret-pod"},
        "responseStatus": {"code": 403},
        "requestURI": "/api/v1/namespaces/default/pods/secret-pod/exec?command=tar&command=cf&command=-&command=/app/secrets",
        "requestObject": null,
        "p_log_type": "Amazon.EKS.Audit"
      }

Detection logic

Condition

not (verb not in ["create", "get"] or resource ne "pods" or subresource ne "exec")
not (responseStatus is_not_null and (responseStatus.code ge "400" or (responseStatus.code ge "1" and responseStatus.code le "16")))
not (username is_not_null and (username in ["masterclient", "aksService"] or (username starts_with "system:" and username not contains "serviceaccount")) and namespace is_not_null and namespace in ["kube-system", "gke-system", "kube-node-lease", "kube-public"])

This rule also runs imperative logic the parser cannot express as a filter; the conditions above are the structured part it could extract.

Exclusions

Top-level NOT(...) conjuncts: predicates this rule actively suppresses.

Field	Kind	Excluded values
`username`	contains	`serviceaccount`
`username`	starts_with	`system:`
`username`	in	`aksService`, `masterclient`
`namespace`	in	`gke-system`, `kube-node-lease`, `kube-public`, `kube-system`
`namespace`	is_not_null	(no value, null check)
`username`	is_not_null	(no value, null check)
`responseStatus.code`	ge	`1`
`responseStatus.code`	le	`16`
`responseStatus.code`	ge	`400`
`responseStatus`	is_not_null	(no value, null check)
`verb`	in	`create`, `get`
`resource`	ne	`pods`
`subresource`	ne	`exec`

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
`username`	contains	`serviceaccount`
`verb`	in	`create` `get`

Output fields

Fields the rule emits when it matches. Chronicle authors list these in the outcome block; they appear on the detection and $risk_score drives alerting. Sentinel / Defender XDR rules build them up through project / summarize / extend stages. Sentinel maps these into alert fields via entityMappings and customDetails; Defender XDR custom detections surface them as alert fields directly.

Field	Source
`username`
`sourceIPs`
`userAgent`
`namespace`
`verb`
`resource`
`requestURI`
`responseStatus`
`cluster`	`p_source_label`
`name`

`j` / `k`	Scroll down / up
`d` / `u`	Half-page down / up
`gg` / `G`	Top / bottom
`h` / `l`	History back / forward
`f`	Follow link (`Shift` = new tab)
`/`	Focus search
`?`	Toggle this help
`↑` / `↓`	Navigate search results
`Enter`	Open highlighted result
`Esc`	Close results / dialog

`type:`	`events` / `rules` / `providers`
`vendor:`	`sigma` / `elastic` / `splunk` / `kusto` / `chronicle` (vendor name alone also works: `sigma:`, `kql:`, `secops:`…)
`tactic:`	TA-id, slug, or name: `credential_access`, `TA0006`
`technique:`	technique or sub-technique ID: `T1003`, `T1003.001` (alias `tech:`)
`severity:`	`critical` / `high` / `medium` / `low` / `informational` (alias `sev:`)
`risk_score`	Numeric comparison on the Elastic risk score (0 to 100): `risk_score>50`, `risk_score<=20`, `risk_score=99` (alias `risk`; Elastic rules only)
`stages:`	Rules with exactly N pipeline stages
`correlation:`	`single_event` / `sequence` / `alternatives` / `alternatives_cross_log` / `all_required` / `correlated`
`with:`	Co-occurrence event-id; stacks (`with:4624 with:4769`) to require all, while a comma list in one occurrence (`with:4624,4769`) is an either-or group. Implies multi-event
`like:`	Structural neighbors of a rule slug (equivalents + subsumption stricter / broader): `like:comsvcs_lsass_memory_dump-splunk-sysmon`
`groupby:`	Entity-grouping substring match against `group_by_keys`: `groupby:user`, `groupby:host`
`uses:`	Rules whose predicate tree touches the field (any kind, any value): `uses:CommandLine`
`excludes:`	Rules with top-level `not()` clauses on the field (FP whitelists): `excludes:ParentImage`
`field:` / `value:`	Predicate search; narrows rule cards to those with a matching leaf and drives the indicator tier. Unquoted = substring, wildcards allowed (`value:mimikatz`)
`indicator:`	Shorthand for `field:F value:V`: `indicator:Image=*\powershell.exe`
`kind:`	Filter by predicate kind. Narrows rule cards to those carrying a matching predicate leaf (`vendor:elastic kind:cidr_match`) and drives the indicator tier: `contains` / `starts_with` / `ends_with` / `regex` / `cidr` / `eq` / `in` … (operator aliases `op:`/`match:`)
`has:` / `no:`	`sample`, `field`, `notes`, `refs`, `trace`, `thirdparty`, `rule`, `pattern`, `timewindow`, `threshold`, `newterms`, `sigma`/`elastic`/`splunk`/`kusto`/`chronicle`
`-op:val`	Exclude matches; works on most operators but not `type:`/`like:`/`has:`/`no:` (use `no:<flag>` to exclude a rule flag): `tactic:execution -vendor:splunk`. Standalone `-kind:`/`-field:`/`-value:` drop every rule carrying a matching predicate leaf (`type:rules -kind:is_null`)
`field:"…"` / `value:"…"`	Quoted value = anchored exact match (also allows spaces): `value:"net user"`
`a,b`	Comma = OR inside one operator (`vendor:sigma,elastic`, `severity:high,critical`); repeating a facet merges the same way. `field:`/`value:` never split (literal commas)
`vendors:` / `stage:`	Singular and plural spellings fold to the canonical operator and value: `tactics:` = `tactic:`, `type:event` = `type:events`, `correlation:sequences` = `correlation:sequence`, `has:thresholds` = `has:threshold`
`"quoted phrase"`	Exact-match a multi-word phrase (free text)