An AWS Security Audit Checklist You Can Run on a Schedule
A repeatable, read-only AWS security audit checklist covering IAM, the root account, S3, security groups, patching, CloudTrail, GuardDuty and billing. Learn how to run it on a schedule and record what you find.
Most AWS accounts drift. Someone creates a test bucket, an old access key stays active, a security group gets opened for a quick test and never closed. None of this is dramatic. It builds up quietly, and a single big audit once a year finds a pile of it at once.
A short, repeatable review works better. This AWS security audit checklist is built to be run on a regular schedule, without changing anything in the account. Every AWS service and feature named below is taken from the official AWS documentation.
Why a cadence beats a one-off audit
A one-off audit gives you a snapshot. By the time you finish fixing the findings, the account has already changed. A regular cadence gives you two things a snapshot cannot.
- Drift becomes visible. When you compare this review to the last one, new users, new keys and new open ports stand out. Without a previous record, everything looks normal.
- Findings stay small. A key that is a few weeks old and unused is easy to explain and remove. A key that nobody has touched for years is a mystery.
Keep the review read-only
An audit that can change things is a risk in itself. If the reviewer can delete a security group or edit a policy, a slip during the review can cause an outage. Keep the review read-only by design.
- Create a dedicated audit role or user that is used only for this review. Do not use an admin identity.
- AWS provides a managed policy called SecurityAudit. Its description says it grants access to read security configuration metadata and is useful for software that audits the configuration of an AWS account. You can attach it to users, groups and roles.
- Note that SecurityAudit includes the permission to generate an IAM credential report. That creates a report file but does not change your resources.
- When you find a problem, do not fix it during the audit. Write it down, then fix it afterwards through your normal change process. This keeps the audit trustworthy and the fixes reviewable.
The checklist
Work through these in the same order each time. The order below starts with the areas where a mistake costs the most.
1. Root account
- Confirm the root user has MFA enabled. The IAM credential report includes an mfa_active column and lists the root user, so you can check it there.
- Check that the root user has no active access keys. The report shows access_key_1_active and access_key_2_active for root as well as for IAM users.
- Look at the last time the root password was used. In the credential report, password_last_used tells you when. If root was used recently and you cannot explain why, note it.
2. IAM users and access keys
The credential report is the fastest way to review this. You can download it from the IAM console under Credential report, or with the AWS CLI commands generate-credential-report and get-credential-report. AWS notes it can be generated as often as once every four hours, which is far more often than you need.
- Every user who signs in with a password should have mfa_active set to TRUE.
- List users who have not signed in or used a key for a long time. Decide whether they still need access.
- Check the age of each access key using the last rotated column, and the last used date.
3. Public S3 buckets
S3 has a feature called Block Public Access. It can be set for your organization, your account, individual buckets and access points, and it has four settings. AWS recommends turning on all four for your account.
- Check that account-level Block Public Access is on. The SecurityAudit policy allows reading this setting.
- List any bucket where Block Public Access is off, and confirm there is a written reason. A bucket that hosts a public website is a valid reason. A forgotten bucket is not.
- Use IAM Access Analyzer for S3 to review buckets that grant public access through ACLs, bucket policies or access point policies. It also reports access shared with other AWS accounts.
4. Security groups and open ports
- List security groups that allow inbound traffic from 0.0.0.0/0 or ::/0.
- Pay closest attention to administrative ports such as SSH (22) and RDP (3389), and to database ports.
- AWS Config has a managed rule named restricted-ssh that flags security groups allowing unrestricted incoming SSH. If you already use AWS Config, read its results instead of checking by hand.
5. Unpatched instances
AWS Systems Manager Patch Manager can scan managed nodes and report only which patches are missing, without installing anything. AWS describes this as a Scan operation, separate from Scan and install.
- Review the patch compliance report in Patch Manager. Note which instances are out of compliance and which patches are missing.
- Keep in mind what compliance means here. AWS says it is measured against the patch baseline you define, and that being compliant does not by itself mean a node is secure.
- List instances that are not managed by Systems Manager at all. These are the ones nobody is patching.
6. Logging with CloudTrail
- Confirm a trail or event data store is logging. CloudTrail is on by default, and event history covers the past 90 days of management events in a Region. It does not show data events, Insights events or network activity events.
- For a longer record, confirm you have a trail or an event data store. AWS says event history is limited to 90 days, and that a trail or event data store is how you keep an ongoing record.
- Use event history to search for recent IAM changes, such as new users, new keys or policy edits, and match each one to a known change.
7. Threat detection alerts
Amazon GuardDuty is a threat detection service. AWS says it analyzes sources such as CloudTrail management events, VPC flow logs and DNS logs, and produces findings.
- Confirm GuardDuty is enabled in every Region you use.
- Read all findings since the last review. For each one, record whether you investigated it, and what you decided.
- Check that findings reach a person, for example through Amazon EventBridge notifications, and not only a console nobody opens.
8. Billing anomalies
Unexpected spend is often the first visible sign of misused credentials or a forgotten resource.
- AWS Cost Anomaly Detection uses machine learning to detect unusual spend and can alert you by email or through an Amazon SNS topic. Confirm your monitors exist.
- Look at recent cost by service and by Region. A service or Region you do not use is worth a question.
How to record findings
A checklist only pays off if the results are written down in the same shape every time. A shared spreadsheet or a simple ticket template is enough. For each review, record:
- The review date and the name of the reviewer.
- For each of the eight areas: pass, needs attention, or not checked, with a short note.
- For each finding: what was found, where it is (account, Region, resource ID), who owns it, and a due date.
- The status of findings from the previous review: closed, still open, or accepted with a reason.
What this checklist does not do
This is a review routine, not a guarantee. Following it does not mean your account is secure, and it does not replace penetration testing or professional advice.
Want a second pair of experienced eyes on your AWS account? Read about our cybersecurity services or our DevOps and cloud services.
Frequently Asked Questions
Need this built?
Book a free 30-minute call. We'll discuss your goals, give you honest advice, and a clear estimate — no obligation.
Ways we can help
MVP Development
Production-ready MVPs in 4-6 weeks. Fixed price, senior engineers.
Learn moreAI App Rescue
Your Lovable, Bolt, or v0 app breaks with real users? We audit, fix the gaps, and hand back code you own.
Learn moreAI Integration
Add AI features to your product — OpenAI, Claude, RAG, semantic search, and copilots.
Learn moreMobile App Development
iOS and Android from one React Native codebase. Fixed price.
Learn moreRelated articles
How to Replace AWS Access Keys with IAM Roles
Long-lived AWS access keys in app config are a standing risk. Here is how to replace AWS access keys with IAM roles for compute and OIDC for CI, with a step-by-step migration plan.
Read moreSecurityYour Supabase App Is Probably Leaking Data — Here's a 2-Minute Check
If you built a Supabase app with Lovable, Bolt, Cursor, or v0 and never wrote Row Level Security policies, anyone can likely read your entire database from a browser. Here's exactly how to check in two minutes — and what to do about it.
Read moreProcessTranslated URLs With English Text: What It Did to Our Indexing
On our own site, blog pages under translated URLs held mostly English text. Here is what the Search Console data showed, what we changed, and what to check on yours.
Read more