What happened
Replaying the AWS console AiTM phishing campaign, 2026
- 1
Look-alike pages and a spoofed notice
Jan 14 – Feb 26, 2026Look-alike domains and a proxy kit are set up, and a spoofed AWS security notice about cross-account IAM role assumption reaches cloud administrators; no source gives a target count.
- 2
Credentials proxied in real time
From February 2026The look-alike page forwards the submitted credentials to the real AWS sign-in in real time and relays the MFA exchange; Datadog says it likely captures one-time codes and has not confirmed it.
- 3
Console session
Date not statedIn one case the attacker signs in to the AWS Console from a VPN exit within 20 minutes of the submission; no source describes what the session did.
Datadog confirms one case in which an attacker signed in to an AWS account with credentials phished through the proxy, within 20 minutes of the submission and from a VPN exit node; it describes no action in the account, gives no victim count, names no victim and no actor, and no source says that any data was read or changed.
Where cloud-provider controls fit
On AWS, a resource control policy on AWS Sign-In refuses console sign-in from outside your trusted networks in the member accounts of an organization, even with valid credentials. A request that a proxy forwards from its own server, or a sign-in from a VPN exit, comes from outside those networks. The sources do not say whether the account in the confirmed case was a member account or which kind of identity signed in. AWS announced these policies on June 16, 2026, after the campaign, and they do not cover console access through IAM Identity Center. Once a session exists, a service control policy on IAM writes and a permissions boundary bound what it can reach, and no source says what this session did. The domains, the email and the password typed into a look-alike page are outside the cloud provider's reach.
Could this have been prevented?
In hindsight, no control would have stopped it. 5 controls would have limited how far it went. 3 steps have no cloud-provider control. The counts below are control assessments, not attack stages: several can address the same step. The same replay with the controls in place before the incident shows where each phase would have been refused or bounded.
- 0
- Prevented
- 5
- Contained
- 3
- Not covered
How it could have been prevented
The same replay with the controls in place. Green marks where a step is refused, blue where it is bounded, and red where no control reaches it. Open a row to read the control, why it acts, and its caveat.
Replaying the AWS console AiTM phishing campaign, 2026
1 · Look-alike pages and a spoofed notice
2 Not covered2 · Credentials proxied in real time
2 Contained1 Not covered6 on other clouds3 · Console session
3 Contained6 on other cloudsKey facts
- Datadog first observed the first kit cluster on February 23, 2026 and says its server was possibly active from the start of February, which is unconfirmed; the domains were registered on January 14 and 15 and on February 25, 2026, and all three share one registrar
- The lure impersonated an AWS security notification that claimed unusual cross-account IAM role assumption patterns in the recipient's AWS Organization; the sender address is treated as spoofed, but Datadog has no original headers, only a partial screenshot
- The kit clones the console sign-in page, mimics the OAuth flow including PKCE parameters and proxies the credentials to the real AWS endpoint in real time, so AWS validates them before the victim is redirected; it relays MFA responses, and Datadog says it likely captures one-time codes and has not confirmed it
- In one case the attacker authenticated to the AWS Console from a Mullvad VPN egress node within 20 minutes of the credentials being submitted; the date, the account and whether an automated pipeline or an operator acted are not stated
- Datadog describes no action after the sign-in; Help Net Security lists potential impact, such as viewing data, modifying resources and creating IAM users, which is not observed activity
- Datadog published on March 9, 2026 and reported the infrastructure to AWS Security, who were coordinating disruption and takedown; AWS announced resource-based policies and resource control policies for AWS Sign-In on June 16, 2026, after the campaign
Sources
Every fact in this breakdown comes from the public reporting below. The replay is a model of that reporting, not a reconstruction of the environment it describes.
- 1.Behind the console: Active phishing campaign targeting AWS console credentialsDatadog Security Labs · Mar 9, 2026 · ResearchLure, kit behavior, infrastructure timeline, MFA relay, the 20-minute sign-in, indicators, detection advice
- 2.Attackers use AiTM phishing kit, typosquatted domains to hijack AWS accountsHelp Net Security · Mar 10, 2026 · NewsTargets as reported, timing as reported, potential impact, mitigations as reported
- 3.Behind the console: An AiTM phishing kit harvesting AWS console credentials and beyondDatadog Security Labs · Jun 24, 2026 · ResearchLater wave, second factors the kit asks for, delivery as inferred
- 4.AWS Multi-factor authentication in IAMAWS · undated · Vendor reportMFA types, passkeys and security keys, phishing resistance, SMS MFA ended
- 5.Controlling console access with resource-based policies and resource control policiesAWS · undated · Vendor reportConsole sign-in restriction, supported actions and condition keys, limits, Identity Center incompatibility
- 6.AWS Sign-in now supports resource-based policies and resource control policiesAWS · Jun 16, 2026 · Vendor reportAvailability date of the sign-in policies, regions, price
The estate in the replay is illustrative and drawn at the level of attack stages; the victim's real environment is not public. Verdicts describe what a control does when it is in place before the incident. They are not an assessment of the victim's security program.