
If your business runs on AWS without a security team, one question keeps coming back: are we exposed, and what should we fix first? This page is for the person who decides: the founder, CTO, head of engineering or IT lead.
AWS Cloud Security is a read-only security review of one AWS account. Your engineer connects it through a role that can look but not touch. It then runs 53 read-only checks across six areas: network exposure, identity, data, servers, logging and stored credentials.
It then links what it finds into attack paths: the step-by-step routes an attacker could take from the internet or an outside account to administrator access or sensitive data. Finally, it ranks the fixes by how many paths each one breaks. It reads configuration only, and it is not a complete security assessment.
Most security tools hand you a long list, each item with a severity. But severity alone does not tell you whether an attacker could use the problem. An unencrypted disk nothing can reach is not the same as an open admin port on a server whose role can do anything.
Breaches rarely come from one dramatic flaw. They come from ordinary misconfigurations that line up: a port left open, a server that hands out its credentials too easily, a role with far more power than it needs. Each looks moderate alone. Together they are a way in.
So AWS Cloud Security reads the list as an attacker would. A finding on an attack path is raised one level; a high or critical finding on no path can be lowered one. The few fixes that matter rise to the top.
Situation. An investor's due diligence or a large customer's security questionnaire asks how well your AWS account is protected.
What it does. A scan of every enabled region gives findings, attack paths and ranked fixes. Your team closes the top items before answering.
What the business gains. Answers based on your real account, and the worst problems closed before anyone asks about them.
Situation. An audit is booked and you want no surprises.
What it does. It finds common problems auditors look for, such as missing multi-factor sign-in, logging switched off and public data. It does not score you against a standard: treat it as preparation, not the audit.
What the business gains. Fewer findings on audit day, and a clear list of what was fixed beforehand.
Situation. A previous agency or team built it, or it came with an acquisition. Nobody can tell you what is in there.
What it does. A first scan shows what is public, which identities are too powerful and what is not being logged. Anything it could not read is listed with the permission it needed, never marked as a pass.
What the business gains. A factual starting point, including what you do not yet know.
Situation. A new product is going live, or an outside company needs access.
What it does. It checks what faces the internet, and flags outside accounts that can reach in and roles that trust another account without a shared code.
What the business gains. No unknown open doors on launch day.
Situation. Something odd happened, or a peer was breached, and the board wants to know where you stand.
What it does. It checks whether the account leaves a trail, such as CloudTrail in every region and threat detection switched on, and flags activity in regions you do not use. It is not an incident investigation.
What the business gains. An honest answer on exposure, and the gaps to close.
Situation. Your engineers change the account every week, so last quarter's review is already stale.
What it does. Each scan is compared with the one before, and new or regressed findings are marked. A finding counts as fixed only when its check ran in full again; otherwise it shows as not re-checked.
What the business gains. Proof that fixes held, and early warning when something slips back.
Situation. Production, staging and old projects sit in separate accounts.
What it does. You connect each account and scan them in turn, one account per scan.
What the business gains. The same standard of review everywhere.
Situation. Long-lived keys and forgotten users are a common way in.
What it does. It flags old access keys, active keys on unused users, root account keys, and people signing in without multi-factor authentication.
What the business gains. Easy-to-miss credential risks found and closed.
Situation. You also want to know about unpatched software and secrets left in settings.
What it does. Two optional scans stay off until you allow them on the role. A vulnerability scan adds known CVEs from Amazon Inspector, missing patches and vulnerable packages, and links vulnerable servers into the attack paths. A secret scan, run by an administrator, finds keys and passwords left in Lambda, EC2 user data, CloudFormation, CodeBuild and ECS configuration, keeping only a fingerprint, never the value.
What the business gains. Patching ordered by real exposure, and leaked keys found before someone else finds them.
Your team can also give an AI agent read access to findings, paths and fixes under your organisation's permissions. A scan an agent asks for runs only with a person's approval.
For the hands-on method behind a review like this, see our guide to running an AWS security review.
Secure by design. It sees only what you allow, and changes nothing. The role uses AWS's read-only SecurityAudit and ViewOnlyAccess policies, with explicit denies on secret values, object contents and decryption. Only Tomorrow Central can use it, with a code tied to your connection. Reports belong to your organisation. Delete the stack and our access ends at once. The rules all our agents follow are in how we secure our AI agents.
Prefer it done for you? Our cybersecurity team can run the review and the fixes.
No. The role it uses cannot create, change or delete your resources. It is always denied secret values, object contents, parameters and decryption, and its own code refuses any call to AWS that is not a read.
Neither. It reads configuration, does not test or attack anything, does not score you against a standard and is not a complete security assessment. It helps you fix the riskiest problems before an audit or a test.
Each scan covers one AWS account, in every region enabled in that account. You can connect several accounts and scan each of them in turn. It does not yet join paths across accounts or scan a whole AWS Organization in one go.
Security Hub runs inside your accounts and checks them continuously against standards, mostly relying on AWS Config; its latest version also shows an attack path graph. AWS Cloud Security is a review you run when you choose, from a read-only role, with nothing to switch on for the configuration review. It ranks fixes by the paths they break and lists what it could not see. You can use both.
Get in touch and we will talk you through pricing for your organisation. If you turn on Amazon Inspector for the vulnerability scan, AWS bills you for Inspector separately.
Delete the CloudFormation stack in your AWS account. Our access ends at once.
Tell us what is coming up: a deal, an audit or a launch. We will set it up with you and go through the first report together. Talk to us.
You can also read the full list of checks on the AWS Cloud Security page.
ITMTB works with enterprises on AI, cloud, security, and digital transformation — from strategy through deployment.