1200kmSECURITY RESEARCH
Loading interactive filters…

1200KM / simulation

T1204.003 Malicious Image — Attack Simulation

Adversaries may rely on a user running a malicious image to facilitate execution. Amazon Web Services (AWS) Amazon Machine Images (AMIs), Google Cloud Platform (GCP) Images, and Azure Images as well as popular container runtimes such as Docker can be backdoored. Backdoored images may be uploaded to a public repository via Upload Malware, and users may then download and deploy an instance or container from the image without realizing the image is…

Complete technique description

Adversaries may rely on a user running a malicious image to facilitate execution. Amazon Web Services (AWS) Amazon Machine Images (AMIs), Google Cloud Platform (GCP) Images, and Azure Images as well as popular container runtimes such as Docker can be backdoored. Backdoored images may be uploaded to a public repository via Upload Malware, and users may then download and deploy an instance or container from the image without realizing the image is malicious, thus bypassing techniques that specifically achieve Initial Access. This can lead to the execution of malicious code, such as code that executes cryptocurrency mining, in the instance or container.[Summit Route Malicious AMIs]

Adversaries may also name images a certain way to increase the chance of users mistakenly deploying an instance or container from the image (ex: Match Legitimate Resource Name or Location).[Aqua Security Cloud Native Threat Report June 2021]

Official technique definition · Open in the interactive matrix

Parent and sub-techniques

Related pages are navigation context. Their tools, rules and lab results are not automatically inherited.

Detection telemetry and configuration

These contracts are selected through exact technique-to-data-component relationships or explicitly labelled planning overrides. Configure a collector at the relevant observation point; a telemetry name alone is not coverage.

Command Execution · source-mapped

DC0064 collection page — Collect process command lines and relevant shell/interpreter audit; shell history alone is neither complete nor tamper-resistant.

Providers and collection configuration

  • Enable Script Block Logging through the policy/configuration for the installed edition. Forward Microsoft-Windows-PowerShell/Operational for Windows PowerShell, or PowerShellCore/Operational for PowerShell 7.
  • Retain script-block ID and fragment ordering alongside process creation. Protect access to collected script text; secrets may be included.
  • On Linux, use narrowly scoped Audit execution rules for the lab account and supported architectures, and retain joined SYSCALL/EXECVE/PATH records.

Required normalized fields: interpreter, command_line, actor_uid, session_id.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0064",
  "collector": "illustrative-lab-collector",
  "observation": {
    "interpreter": "/bin/sh",
    "command_line": "printf LAB_TELEMETRY_CHECK",
    "actor_uid": 1000,
    "session_id": "lab-session-1"
  }
}

Visibility limits: Script content and process command lines are different signals. Logging may be partial, fragmented, filtered or disabled; never assume a command line reveals everything executed.

Container Creation · source-mapped

DC0072 collection page — Collect runtime create events and image/container IDs; correlate with API audit to establish the initiating identity.

Providers and collection configuration

  • Subscribe to the lab daemon event stream with filters for the relevant object/action; send it to a durable collector.
  • Record runtime/container ID, image digest, action, timestamp and host. Join orchestrator identities where available.
  • Use a protected collector connection. Docker socket access is privileged; never expose it publicly or grant it to the website.

Required normalized fields: container_id, image_digest, action.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0072",
  "collector": "illustrative-lab-collector",
  "observation": {
    "container_id": "lab-container-1",
    "image_digest": "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
    "action": "create"
  }
}

Visibility limits: Runtime event history is bounded and can be lost without continuous collection. Image tags are mutable; pin digests. A lifecycle event is not a full process or file audit.

Container Start · source-mapped

DC0077 collection page — Collect runtime start events and join creation, image and later process activity. An orchestrator request alone is insufficient.

Providers and collection configuration

  • Subscribe to the lab daemon event stream with filters for the relevant object/action; send it to a durable collector.
  • Record runtime/container ID, image digest, action, timestamp and host. Join orchestrator identities where available.
  • Use a protected collector connection. Docker socket access is privileged; never expose it publicly or grant it to the website.

Required normalized fields: container_id, action, runtime.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0077",
  "collector": "illustrative-lab-collector",
  "observation": {
    "container_id": "lab-container-1",
    "action": "start",
    "runtime": "lab-docker"
  }
}

Visibility limits: Runtime event history is bounded and can be lost without continuous collection. Image tags are mutable; pin digests. A lifecycle event is not a full process or file audit.

Image Creation · source-mapped

DC0015 collection page — Collect VM image create/register APIs from the cloud control plane; retain source instance/disk, image ID and final state. Container image events are a separate concept.

Providers and collection configuration

  • In the isolated account/project/subscription, route administrative audit logs to a protected destination with retention and access controls.
  • Select the relevant service, regions and read/write operations. Confirm whether the API below is a management or data event in the provider documentation.
  • Preserve actor/session, resource, request ID, service, operation, region and error details. Check delivery latency and avoid logging credentials or sensitive request bodies.

Required normalized fields: image_id, source_instance, action, state.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0015",
  "collector": "illustrative-lab-collector",
  "observation": {
    "image_id": "image-lab-1",
    "source_instance": "vm-lab-1",
    "action": "create",
    "state": "pending"
  }
}

Visibility limits: AWS, Azure and Google expose different APIs and event schemas. An accepted API request does not prove the final resource state; join asynchronous outcomes and inventory.

Instance Creation · source-mapped

DC0076 collection page — Collect launch/create request and completion state, retaining instance ID, image, region and initiating principal.

Providers and collection configuration

  • In the isolated account/project/subscription, route administrative audit logs to a protected destination with retention and access controls.
  • Select the relevant service, regions and read/write operations. Confirm whether the API below is a management or data event in the provider documentation.
  • Preserve actor/session, resource, request ID, service, operation, region and error details. Check delivery latency and avoid logging credentials or sensitive request bodies.

Required normalized fields: instance_id, image_id, action, state.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0076",
  "collector": "illustrative-lab-collector",
  "observation": {
    "instance_id": "vm-lab-1",
    "image_id": "image-lab-1",
    "action": "create",
    "state": "provisioning"
  }
}

Visibility limits: AWS, Azure and Google expose different APIs and event schemas. An accepted API request does not prove the final resource state; join asynchronous outcomes and inventory.

Instance Start · source-mapped

DC0080 collection page — Collect start APIs and instance-state transitions with request IDs; distinguish accepted from completed.

Providers and collection configuration

  • In the isolated account/project/subscription, route administrative audit logs to a protected destination with retention and access controls.
  • Select the relevant service, regions and read/write operations. Confirm whether the API below is a management or data event in the provider documentation.
  • Preserve actor/session, resource, request ID, service, operation, region and error details. Check delivery latency and avoid logging credentials or sensitive request bodies.

Required normalized fields: instance_id, action, state_before, state_after.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0080",
  "collector": "illustrative-lab-collector",
  "observation": {
    "instance_id": "vm-lab-1",
    "action": "start",
    "state_before": "stopped",
    "state_after": "pending"
  }
}

Visibility limits: AWS, Azure and Google expose different APIs and event schemas. An accepted API request does not prove the final resource state; join asynchronous outcomes and inventory.

Network Traffic Content · source-mapped

DC0085 collection page — Scope approved packet capture/protocol content collection to the lab flow; retain capture provenance and truncation indicators.

Providers and collection configuration

  • Zeek: Connection and protocol metadata from traffic visible to the sensor.
  • Suricata EVE: Configured flow, alert, DNS, HTTP, TLS and protocol records.
  • Use an authorized lab TAP/SPAN, virtual mirror or gateway interface. Define which traffic crosses it; avoid assuming visibility into every segment.
  • Enable the required Zeek analyzers or Suricata EVE event types. Export logs with sensor identity, clock synchronization and flow correlation identifiers.
  • Monitor capture loss, truncation and exporter sampling. Capture payload only with approval and restrictive retention; encryption normally prevents plaintext inspection.

Required normalized fields: flow_id, protocol, method, path, payload_excerpt.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0085",
  "collector": "illustrative-lab-collector",
  "observation": {
    "flow_id": "lab-flow-1",
    "protocol": "http",
    "method": "GET",
    "path": "/lab-check",
    "payload_excerpt": "LAB_TELEMETRY_CHECK"
  }
}

Visibility limits: Flows are not packet payloads. NAT, asymmetric routes, encryption, missing mirrors and sampling can hide attribution or content. A destination connection alone does not prove malicious intent.

Process Creation · source-mapped

DC0032 collection page — Collect Sysmon ProcessCreate (1); preserve ProcessGuid and parent identifiers. Security 4688 is an alternative with separate audit/command-line settings.

Providers and collection configuration

  • Microsoft Sysmon: Windows event-based collection; enable the event types needed below.
  • Linux Audit: Linux alternative for supported system-call and file events; different semantics and fields.
  • Apple Endpoint Security clients: macOS alternative where the subscribed event exists; requires an entitled, approved client, not an iOS collector.
  • On a disposable Windows host, review the installed Sysmon schema and current configuration; merge scoped event filters into the existing policy rather than replacing it.
  • Forward Microsoft-Windows-Sysmon/Operational to the lab collector. Preserve event ID, timestamp, computer, process identifiers and the original event.
  • For Linux or macOS, select the equivalent sensor separately and test its emitted fields; a Windows event ID does not transfer across platforms.

Required normalized fields: process_guid, image, parent_image, command_line, user.

<EventFiltering>
  <ProcessCreate onmatch="include">
    <Image condition="is">C:\Windows\System32\whoami.exe</Image>
  </ProcessCreate>
</EventFiltering>

This fragment only selects the named process. It is not a complete production policy. Inspect the installed schema with sysmon64 -s; preserve existing rules, then apply the reviewed complete configuration with sysmon64 -c <configuration-file>.

Illustrative collection event

Project-normalized synthetic JSON, not a native provider log, captured event, attack verdict or validated detection. A parser/adapter is required for a real SIEM. Example names, IPs and values are fictional.

{
  "schema": "1200km.telemetry.example.v1",
  "synthetic": true,
  "timestamp": "2026-09-27T12:00:00Z",
  "telemetry_id": "DC0032",
  "collector": "illustrative-lab-collector",
  "observation": {
    "process_guid": "lab-process-001",
    "image": "C:\\Windows\\System32\\whoami.exe",
    "parent_image": "C:\\Windows\\System32\\cmd.exe",
    "command_line": "whoami",
    "user": "LAB\\analyst"
  }
}

Visibility limits: Event filtering and sensor versions change coverage. High-volume image-load and process-access collection needs tuning; endpoint events alone do not establish intent.

Detection rules and analytic guidance

Open the dedicated detection workspace · Existing query engineering

DET0248 User Execution – Malicious Image (containers & IaaS) – pull/run → start → anomalous behavior (T1204.003)

MITRE detection strategy

AN0691 Analytic 0691

CONTAINERS (Docker/K8s/containerd): A user pulls an untrusted image from a public/unknown registry and then creates/starts a container from that image. Shortly after start, the container spawns unexpected utilities (e.g., curl/wget/bash/python), or makes outbound network connections atypical for the namespace/workload. The analytic correlates Image Creation/Download → Container Creation → Container Start → Command Execution/Network activity within a short window and with a consistent image digest.

Platforms: Linux.

  • Image Creation — containerd:events; channel Image pull from untrusted registry (name NOT IN allowlist) or new digest never seen before.
  • Container Creation — kubernetes:audit; channel create: Pod/Container created with image tag 'latest' or mutable tag; imagePullPolicy=Always; noDigest=true.
  • Container Start — kubernetes:events; channel start: ContainerStarted or Pulling image → Started container.
  • Command Execution — auditd:SYSCALL; channel execve: Process in container namespace executes curl|wget|bash|sh|python|nc with outbound args.
  • Network Traffic Content — NSM:Flow; channel New egress from container IP/namespace to Internet or non-approved CIDRs/ASNs.

AN0692 Analytic 0692

IAAS (Cloud images/VMs): A new VM/instance is launched from a non-approved or newly-seen image (AMI/GCP Image/Azure Image). On first boot, cloud-init/user-data or embedded agents download code, spawn system utilities, or open outbound C2/mining traffic. The analytic correlates Instance/Image Creation → Instance Start → in-guest Process/Command Execution and/or anomalous network traffic.

Platforms: Windows.

  • Instance Start — AWS:CloudTrail; channel RunInstances.
  • Instance Creation — azure:activity; channel Microsoft.Compute/virtualMachines/write: imageReference publisher NOT IN allowlist OR plan is new/unknown.
  • Process Creation — WinEventLog:Sysmon; channel EventCode=1.
  • Network Traffic Content — NSM:Flow; channel New VM egress to crypto-mining pools or non-approved Internet ranges within minutes of boot.
Source rule directory and observation requirements

No exact-tagged Sigma source rule was imported. This does not establish that the technique is undetectable; the MITRE guidance and collection requirements remain relevant.

Before treating a rule as coverage

  1. Read the actual condition and logsource; identify attacker-side, victim-side or third-party visibility.
  2. Map native sensor fields, parse times and identities, then compile for the selected backend.
  3. Replay a behavior-positive case, a normal control, an authorized look-alike and a missing-telemetry case.
  4. Record matches and misses with rule version and source events. No match is inconclusive when required telemetry is absent.

Anomaly-based detection

No exact technique-specific Atlas model is present for T1204.003 in the selected source snapshot. The following is a design worksheet, not a newly validated model.

Anomaly engineering worksheet

Start from the exact analytic above. The collection-specific lenses below are analyst recommendations; they do not assert that an arbitrary statistical deviation detects this technique.

  • Script and command execution: Per host role and principal: rare interpreters, command argument patterns and parent-child sequences. Normalize scripts carefully; obfuscation, administrative automation and missing command lines change visibility.
  • Container runtime: Per workload image and namespace: rare process starts, unexpected image/privilege changes and new network relationships. Use workload/deployment identity, not short-lived container IDs alone.
  • Cloud control plane: Per account, role, region and API: rare actions, new role-resource relationships, enumeration bursts and configuration changes. Separate deployment automation and organization-level administrative activity.
  • Network and protocol telemetry: Per source and observation point: connection rate, destination-host/port fan-out, response/failure ratio and previously unseen destinations. Compare equivalent time windows; separate authorized scanners, NAT and sampling effects.
  • Endpoint activity: Per host role and user: rare process/object combinations, bursts of object access or modification, and ordered parent/action sequences. Compare maintenance windows and signed software rollout activity.
  1. Define an entity/peer group and feature; exclude the evaluation period from baseline training.
  2. Use representative normal and maintenance activity. Do not invent a production threshold from a synthetic example.
  3. Evaluate held-out behavior-positive and benign cases; measure false positives, misses and telemetry loss separately.
  4. Keep abnormality separate from maliciousness. Correlate with the behavior-specific source evidence.

Anomaly Detection Atlas · Baseline, validation and blind spots

Attack tools and documented software

No reviewed lab-tool association is selected for this exact technique. This is a curation gap, not a statement that no tool can implement the behavior.

Simulation design and documented procedures

Atomic evidence: no compatible candidate in this snapshot. Live lab status: not run for these imported procedures.

Use disposable lab hosts and test accounts, explicit target allowlists, a clean snapshot, bounded egress and a restore plan. Review every candidate and dependency before any execution.

No compatible Atomic procedure is present in this pinned evidence set. Technical feasibility has not been ruled out. A custom procedure must define the behavior, target, collection point, success observable and rollback before execution.

Technique-specific experiment contract

  1. Choose one procedure implementing T1204.003 Malicious Image; identify the observable behavior in the source definition, not just its tactic.
  2. Record exact target, tool/version, inputs, privileges, network limits and approval. Never run an imported default target automatically.
  3. Enable the collection sources above and verify a benign baseline reaches the detector.
  4. Run only the approved isolated scenario; capture native events and an independent behavior observable.
  5. Evaluate the exact rule/model, restore the lab, and preserve configuration, event hashes, alerts, controls and misses.

Synthetic logs and offline contract check

Download synthetic collection fixture (JSON) · Download the complete engineering workbook

This fixture exercises the listed collection fields. It does not reproduce Malicious Image, label its example events as malicious, or measure detection effectiveness. Use the exact analytic and selected procedure to build an independently labelled behavior-positive dataset.

Not run in this browser. No network scan or attack command is executed by this control.

Validation status and remaining work

Source identity
Joined by exact technique/object IDs against hash-pinned sources.
Collection fixture
Available for offline schema exercises; not attack emulation.
Native telemetry / backend / real attack
Not validated for these imported procedures. Source fidelity and page tests cannot establish sensor or detector effectiveness.

All-page audit and evidence inventory · This page’s machine-readable workbook

Sources and evidence boundaries

Pinned MITRE STIX source · SHA-256 81113ba98afa2928f3dcff62022539f7ce30c315c32e14422f11ff95ce534352. Definition and procedure context are reproduced from this snapshot, not independently re-investigated incidents.

Source-linked mitigations

Connected ecosystem references

Linked tags

Detection and collection

T1204.003 detection workspace

No reviewed association in this snapshot.

Threat actor context

These are explicit actor-to-technique associations in the existing Threat Matrix snapshot, not attribution of an event or proof that a detector identifies the actor. No tool-to-actor relationship is inferred.

Existing research

Threat Matrix: knowledge routes, evidence and actor context

Pinned research references. No browser attack runner or production-validated detector is asserted. Imported procedures remain unvalidated; any bounded lab evidence has its own scope. ATT&CK / Atomic provenance · Detection provenance.