Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
A public S3 bucket is cloud storage that anyone on the internet can read, and sometimes write, without credentials. It remains one of the most reliable sources of large data breaches.

What Is a Public S3 Bucket?
A public S3 bucket is an Amazon storage bucket whose contents can be listed or downloaded by anyone on the internet, with no AWS account and no credentials. It happens through a bucket policy with a Principal of "*", through a legacy access control list granting the AllUsers group, or through an object-level permission that survives a tightened bucket. The same mistake exists on every platform — Azure containers set to blob or container access, Google Cloud buckets granting allUsers — but S3 has the longest and most public record of it.
The causes are mundane and repeat across organisations. A developer makes a bucket readable to serve static assets, then later uses the same bucket for exports. A third-party supplier provisions storage on your behalf with permissive defaults. A troubleshooting session opens access “temporarily” to confirm the network is not at fault, and the change is never reverted. Automation copies an existing template, inheriting a policy nobody has read since it was written.
One misreading deserves its own mention. The access control list group named AuthenticatedUsers does not mean users of your account. It means any AWS account holder anywhere, which is effectively the public with one extra step.
Public buckets have produced some of the largest avoidable exposures of the past decade. In 2017 a supplier working for a US telecommunications firm left records for roughly 14 million customers readable, and separate research found an exposed consultancy bucket containing credentials and key material. A writable bucket is worse still: anyone who can upload gains a file on your domain, which is enough to serve malware, host a phishing page under your brand or pass a domain-ownership check. In 2022 an airline left approximately 6.5 TB of flight data, source code, and crew information open. The pattern never changes: the data is discovered by a researcher or an opportunist scanning common bucket names, not by the owner.
Write access is worse than read access. An attacker who can write to a bucket serving your website can replace a JavaScript file, which turns a storage misconfiguration into a supply chain compromise for every visitor.
Amazon made meaningful default changes in April 2023: new buckets have Block Public Access enabled and object access control lists disabled unless you deliberately turn them back on. Older buckets did not change, which is where the remaining risk sits.
IAM Access Analyzer reports every bucket reachable from outside your account or organisation, including access granted through policy conditions that are harder to read by eye. Trusted Advisor and the S3 console both flag public buckets, Amazon Macie identifies which of them contain personal or financial data, and open-source tools such as Prowler and ScoutSuite produce the same findings on demand. Continuous coverage belongs to cloud security posture management, and the same checks should run against Terraform or CloudFormation before deployment rather than after.
Logging deserves attention too. Server access logging or a CloudTrail data event trail on the bucket is what later tells you whether anyone actually downloaded the contents, which is the first question your legal team will ask.
If a public bucket held personal data, you have a personal data breach whether or not you can prove anyone fetched it. The controller must assess the risk and, where notification is required, tell the ICO within 72 hours of becoming aware, with notification to affected individuals where the risk to them is high. Being unable to show who accessed the data generally pushes the assessment towards the cautious end, which is a practical argument for enabling access logging on anything holding personal data.
A public bucket is one instance of a broader category, covered in our entry on security misconfiguration. If you want an independent check across your estate, our cloud security services include exposure review as standard.
Mostly by guessing names and checking them at scale. Bucket names are globally unique and predictable, so wordlists built from company names, product names, and common suffixes such as backup, dev, logs, and assets find a great deal. Certificate transparency logs, JavaScript files on public websites, mobile application binaries, and search engines indexing bucket listings supply the rest.
It blocks the direct public-bucket approach, which is the point. The supported pattern is to keep the bucket private and put CloudFront in front of it using origin access control, so only the distribution can read the objects. You gain HTTPS with your own certificate, caching, request logging, and the ability to attach a web application firewall, none of which a public bucket provides.
Not by itself. Removing public access leaves the question of which identities inside your organisation can read it, and broad policies granting a whole account access are common. Apply least privilege in the bucket policy, use VPC endpoint conditions to restrict the network path, and enable default encryption with a key whose policy adds a second check on who may decrypt.
Close the access first, then preserve evidence rather than deleting anything. Pull server access logs and CloudTrail data events to establish what was retrieved and by whom, identify the data categories and the number of individuals affected, and start the 72-hour clock for assessing notification to the ICO. Treat it as an incident with a documented timeline, not as a configuration fix.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.