Over 106 million individuals have had their credit application details stolen from Amazon S3 data buckets which had been insecurely configured by Capital One.

The 30Gb of data from the S3 bucket was downloaded and then posted onto a Github account by the attacker. The stolen data was soon spotted and then reported to Capital One’s responsible disclosure email account.

The FBI has arrested systems engineer Paige A. Thompson and charged her with the breach. Her twitter account (now suspended) included several descriptions of how she breached multiple Amazon AWS environments that had not been correctly configured and secured.

Our article explaining What is a supply chain attack describes the danger of poorly configured cloud servers which could allow attackers to access your data or application code.

The relative ease of establishing a new cloud server on cloud services such as Amazon Web Services or Microsoft Azure may lead to a false sense of confidence that it is similarly easy to correctly secure that cloud environment.  Wise network administrators will seek the independent validation of their cloud security by engaging an expert to perform a network penetration test.

Amazon offer a best practice guide to help their clients secure S3 resources

UPDATE 9th August 2019:

Further research on this attack has identified that the initial attack vector was a Server Side Request Forgery against the Web Application Firewall protecting the AWS environment used by Capital One.  A detailed description of the SSRF attack has been published by one of CloudFlare’s security team.

Honda leaks 134 million security records

In related news – Honda had mistakenly published an unsecured Elasticsearch database to the internet which contained live details of employee computer systems including their patch levels and whether endpoint security protection was active.

This database contained 134 million rows and included details of employees’ systems allowing would be attackers to target the PC of officers such as the CEO, CFO directly by IP address.

Knowing the exact software version running on a given computer, including which patches had been applied and versions of the security software on the endpoint, makes the job of an attacker much easier when trying to gain access to that device.

The database was discovered using a Shodan search by a security researcher, who reports that it then took him two days to locate a contact in Honda’s security team in order to report the problem to them. In the end he had to resort to Twitter to ask for contact details.

Security Managers should consider how to make it easy for people to responsibly inform them of breaches and leaks through security@ email accounts or contact forms on their website.