D014F3
Compute
🖥️ Honey Container Image
Decoy images in ECR/ACR/GCR registries.
What it is
A decoy container image in a registry, with a realistic repository name and tag history and honeytokens inside its layers. Attackers pull images to look for embedded credentials and source code. The pull is logged by the registry, and any honeytoken taken from the image alerts again when it is used.
Implementation requirements
- Push a decoy image to ECR, ACR or Artifact Registry with a realistic repository name and tag history. Use a real base image, not an empty one
- Monitoring: AWS records image pulls as
BatchGetImageandGetDownloadUrlForLayerin CloudTrail. Azure:ContainerRegistryRepositoryEventsvia a diagnostic setting. GCP: Artifact Registry Data Access logs (enable DATA_READ) - Layer a second tripwire inside the image: honeytokens (D006) in environment files or config, which alert if someone runs the image and uses them
- Safety: the image contains no real code, credentials or internal hostnames; no workload is configured to pull it
- OPSEC: tags, push dates and size consistent with neighbouring repositories
- Rollback: Delete the image and repository. Revoke any honeytokens baked into its layers.
- Upkeep: Push new tags at a cadence like real repositories. Quarterly proof-of-life pull.
Technologies
AWSAzureContainerGCP
ATT&CK coverage 6
| Technique | Tactic | Status | Confidence | Tier |
|---|---|---|---|---|
| T1204.003 Malicious Image |
Execution | theoretical | weak | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (1 source)
No source shows a decoy observing a user running a malicious image. | ||||
| T1610 Deploy Container |
Execution | confirmed | strong | T3 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (3 sources)
| ||||
| T1195 Supply Chain Compromise |
Initial Access | theoretical | weak | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (1 source)
The compromise happens upstream of the defender. | ||||
| T1195.002 Compromise Software Supply Chain |
Initial Access | theoretical | weak | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (1 source)
The compromise happens upstream. | ||||
| T1525 Implant Internal Image |
Persistence | theoretical | moderate | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (2 sources)
| ||||
| T1612 Build Image on Host |
Stealth | theoretical | moderate | T2 |
| ↳ Detection point: seen once the attacker is working inside a high-fidelity decoy | ||||
↳ Evidence (2 sources)
| ||||
Evidence
- Decoy Object (D3-DO) — MITRE D3FEND standard
- MITRE ATT&CK Enterprise v20.0 — MITRE 2026 standard
- 2023 Honeypotting in the Cloud Report — Orca Security 2023 industry
- GitGuardian Honeytoken — core concepts — GitGuardian industry