What happened
Replaying the SES-pionage intrusion / dates not stated; Permiso investigated in October 2022
- 1
A key in a public repository
Not stated; no later than October 2022A long-lived AWS access key and its secret key sit in a public GitHub repository that anyone can read; who placed them there, when and for how long is not stated.
- 2
The exposed key at the AWS API
Not stated; no later than October 2022Someone signs calls to the client's AWS account with the exposed key; who it is, where the calls come from and whether the key was taken from the repository are not stated.
- 3
Querying Amazon SES
Not stated; no later than October 2022The key calls three actions that show the SES sending quota, the sandbox status and the configured sender identities; whether the calls succeed is not stated, and Permiso reports no mail sent.
The key was used to call actions that show the Amazon SES sending quota, the sandbox status and the configured sender identities; whether the calls succeeded is not stated. Permiso reads such calls as a sign that the account may be tested or targeted for mail abuse, and reports no mail sent and no later action.
What an exposed key can learn about your Amazon SES
A long-lived access key and its secret sat in a public GitHub repository, and the key was used to query Amazon SES's sending quota, sandbox status and sender identities; Permiso reports no mail sent. What you control is what such a key can do before anyone finds it. A service control policy can refuse new long-term access keys outside approved administrators, refuse IAM users' calls that carry no MFA, allow calls only from your own networks, and deny Amazon SES in accounts that send no email; each holds only under its own precondition, and the quota call falls outside the SES deny. GitHub push protection can refuse a push that carries the key, unless the pusher can bypass it. What a repository holds is decided in GitHub, not in your AWS account.
Could this have been prevented?
In hindsight, no control would have stopped it. 6 controls would have limited how far it went. 1 step has 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
- 6
- Contained
- 1
- 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 SES-pionage intrusion / dates not stated; Permiso investigated in October 2022
1 · A key in a public repository
2 Contained1 Not covered3 on other clouds2 · The exposed key at the AWS API
2 Contained3 on other clouds3 · Querying Amazon SES
2 ContainedKey facts
- Permiso investigated an alert at an unnamed client in October 2022 and traced it back to a long-lived AWS access key and its secret key exposed on a public GitHub repository
- The write-up did not say who placed the key in the repository, when, for how long it was public, or which IAM identity it belonged to; that the key's user took it from the repository is unconfirmed
- Among many findings, Permiso singled out three calls made with the key, which show the SES sending quota and sandbox status and list the email addresses and domains configured as senders; their dates, their source and whether they succeeded were not stated
- Permiso reported no email sent, no sender identity created or verified and no IAM change after the three calls
- Permiso noted that an account in the SES sandbox can send at most 200 messages per 24 hours and one per second, and only to verified addresses and domains, and said such calls by a long-term key may signal that an organization is being tested or targeted for mail abuse
- Permiso published its account, SES-pionage, on January 12, 2023; its response advice was reactive and named no preventive AWS control
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.SES-pionagePermiso · Jan 12, 2023 · ResearchKey exposure on GitHub, the SES calls made with the key, investigation month, SES sandbox background, response advice
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.