Profiling Sea Turtle: Tactics, History & Defenses

20.8.2025

De blogs zijn enkel beschikbaar in het Engels.

Summary

  • Threat Actor: Sea Turtle (Türkiye-nexus), a.k.a. Teal Kurma, Cosmic Wolf, Marbled Dust
  • Motivation: State-affiliated espionage
  • Targeting: AWS and/or Azure environments, Government, Internet Service Providers, Technology, Telecommunications, etc.
  • Attack Types: DNS-hijacking, zero-day exploitation, credentials and cloud storage theft, etc. 
  • Defenses: Identity security, storage protection, network traffic monitoring, vulnerability management, etc.
If you enjoy reading about cloud threat actors, you might also enjoy our training or investigating your own cloud incidents with Cloud Labs.

Overview

Sea Turtle is a Türkiye-nexus threat actor known for conducting state-affiliated espionage operations since at least 2017. The group has primarily targeted government agencies, IT service providers, and the telecommunications sector, with a notable evolution in tactics over the past decade.

Sea Turtle initially gained attention for its DNS hijacking campaigns, which redirected traffic to steal credentials, often targeting Middle Eastern and European entities. This activity persisted from 2017 through 2020. By 2021, the group had shifted toward exploiting exposed credentials and known vulnerabilities to gain access to internet service providers and related infrastructure, often as a path to reach downstream victims. This mentioned activity continued into 2023, when Sea Turtle was observed deploying a custom Linux/Unix implant and expanding into cloud-focused intrusions. 

In one case, the group targeted a technology company’s cloud environment using stolen credentials to attempt data theft. Using the cloud environments command line they altered security group settings to allow direct Secure Shell (SSH) access. Since April 2024, Sea Turtle has been observed exploiting unpatched user accounts vulnerable to a zero-day (CVE-2025-27920) in Output Messenger, a cross-platform enterprise messaging application.

Cloud TTPs

All of the details below are also available in our GitHub repository, which provides background on each observed technique along with corresponding MITRE ATT&CK mappings, making it easy to map these insights directly to your cloud logs.

Sea Turtle – Tactics Accordion

Click a tactic to view the procedure details.

Tactic
ID
Technique
Initial Access
T1078.004
Valid Accounts: Cloud Accounts
Sea Turtle used stolen AWS IAM keys or Azure Entra ID credentials to authenticate and access cloud environments, enabling unauthorized API calls.
Execution
T1059.009
Command and Scripting Interpreter: Cloud API
Executed malicious commands via AWS CLI or Azure CLI/PowerShell to manage resources (EC2/S3/IAM; VMs/Blob/Entra ID).
Persistence
T1543.005
Create or Modify System Process: Container Services
Modified security group rules to allow SSH (22) from attacker infrastructure, ensuring persistent access.
Defense Evasion
T1578
Modify Cloud Compute Infrastructure
Opened SSH by altering network controls, blending with legit admin activity to evade detection.
Collection
T1530
Data from Cloud Storage
Queried or downloaded sensitive data from cloud storage/compute using stolen credentials and CLI access.
Command and Control
T1071
Application Layer Protocol
Used SSH (TCP/22) as the C2 channel to manage compromised instances remotely.
Exfiltration
T1041
Exfiltration Over C2 Channel
Likely transferred data to actor infrastructure through the established SSH-based C2 channel.

Inferences and Limitations

In CrowdStrike’s 2023 Cloud Risk Report, they briefly describe how Cosmic Wolf (a.k.a. Sea Turtle), targeted data in a victim’s cloud environment. The report is provider-agnostic, yet the combination of ‘security group’ changes and CLI usage suggests either AWS or Azure.

Additionally, we assess it is likely the actor used T1530 (Data from Cloud Storage Object) and T1537 (Exfiltration to Cloud Account), inferred from the objective of targeting stored cloud data; this is consistent with common cloud-intrusion tradecraft. However, the report does not provide details on:

  • How credentials were stolen;
  • What specific data was accessed (e.g., S3 vs. RDS); or
  • Whether any exfiltration was confirmed.

Furthermore, there is no mention of the use of other services (e.g., Lambda, IAM role creation, or temporary credentials), which limits visibility into potential persistence or privilege escalation techniques (e.g., T1078.004, T1098).

Incident Response Steps

Below are several high-impact steps to prepare for and respond to the TTPs mentioned above. The list isn’t exhaustive, but it highlights key areas where defenders can focus their efforts. Click each step to view additional details.

Harden
  • Rotate compromised keys using AWS SSM or CLI.
  • Transition IAM users to federated roles via AWS SSO to prevent recurrence.
  • Rotate compromised credentials (reset passwords, revoke app secrets) using Entra ID portal or PowerShell.
  • Transition to managed identities or federated roles via Entra ID to prevent recurrence.

Getting Help

If you spot any of these TTPs or indicators in your environment, or just want to become incident-ready in the cloud, contact our team. We’ll help you tighten your AWS defenses, refine your IR runbook, and stay ahead of Sea Turtle or other cloud-focused adversaries.

About Invictus Incident Response

We are an incident response company and we ❤️ the cloud and specialize in supporting organizations in preparing and responding to a cyber attack. We help our clients stay undefeated!

🆘 Incident Response support reach out to cert@invictus-ir.com or go to https://www.invictus-ir.com/24-7

Meer nieuws