The Cyber Empowerment Hub Ltd
Security
As a security company, we take our own security incredibly seriously. That's why we have implemented enterprise-level security tooling to help secure us and your data.
The surest way to keep information safe is not to hold it.
So the list of what we keep is deliberately short, and the list of what we refuse to keep is the more useful one. Both are set out in full in our privacy notice.
Five things, and that is the whole list
- The email addresses on your plan. Yours, or the ones you added for your team. Without them we cannot send the practice emails, which is the service.
- What the practice showed. Which emails went to whom and what happened next. That is what your training report is made of.
- Your subscription record. Which plan, how many seats, and Stripe's reference numbers, so we know who has paid.
- Anything you send us. If you email us, we keep the conversation while it is useful, then delete it.
- Anything forwarded to the scam checker. Held while we look at it, then deleted automatically after 30 days, along with our record of it.
The parts worth stealing, we do not have
- No passwords, ever. Our practice pages only teach. They never ask anyone for a password and never record what is typed.
- No card numbers. Payment happens on Stripe's own checkout page. We never see a card.
- No tracking. This site sets no cookies and runs no analytics, so we do not know you visited unless you tell us.
- Nothing kept out of habit. Plan data goes within 90 days of you cancelling. Billing records stay six years because tax law says so, and that is the only reason.
- Nothing sold. Not to advertisers, not to anyone, at any price.
We do not own a server room. The service runs on Amazon Web Services in Ireland, so your data stays in Europe, and the parts a small company would be worst at are handled by a company that does nothing else.
Amazon looks after the buildings, the locked doors, the power and the hardware. We look after everything we put on top, and that half is checked every morning.
Shared responsibilityAWS eu-west-1, Ireland. Physical and platform security sits with the provider. Configuration, access and data handling are ours, and are audited daily.
Every setting in our Amazon account is compared each morning against several hundred published rules for how a cloud account ought to be set up.
Daily account auditAutomated scan of the live account against recognised industry benchmarks. Catches the gap between what was intended and what is actually there, which written policy never closes on its own.
Nothing in our build system holds a permanent key to that account. Each job gets a temporary pass that expires within the hour and opens only the one door it needs.
Short-lived credentialsFederated identity per run, scoped to a single repository, expiring inside the hour, carrying only the permissions that job requires. No long-lived access key exists to be stolen.
The account keeps a permanent log of every action taken inside it, and the tool that inspects it can read settings but cannot read the contents of anything.
Audit trail, read-only auditorAccount-wide activity logging to a separate store. The auditing role holds configuration read permissions and no object or record access at all.
Software goes wrong in a handful of predictable ways.
Each check below covers one of them. They all run on their own schedule, on every change we make and again every morning, and a person reads what they found the same day.
A machine reads every line of code we have written, looking for the small set of mistakes that cause most break-ins.
Code scanning · SASTEvery proposed change, plus daily. Injection, unsafe comparison, path traversal, weak crypto.
Most modern software is mostly other people's software, so every outside part we use is checked against the public lists of known faults.
Third-party dependencies · SCAEvery change, plus daily. Matched by ecosystem and exact version rather than by guessed product name, including the nested ones.
Our whole history is searched for passwords and keys typed in by mistake, and anything found is then tested to see whether it still works.
Secret scanningPattern scan on every change, full history walk daily, with live verification against the issuing provider. A dead string is not treated as an incident and a live one is not treated as noise.
Our servers, storage and databases are written down as files rather than clicked together by hand, so a change can be read and argued with before it is switched on.
Infrastructure as codeThe whole estate is declared. Every change is checked against a large policy library for public exposure, over-broad permissions and missing logging, before it is applied.
The live website is tested from the outside every day, the way a visitor reaches it, because code can be right and the thing on the internet still wrong.
Website testing · DASTDaily passive scan of the public sites: security headers, cookie handling, exposed paths, plus a sweep against a library of thousands of published weaknesses.
The rough version of that test, the one that behaves like an attacker, runs against a private copy of the site, so nothing a customer is doing is ever interrupted.
Active scanning, isolatedThe full attack-style pass runs against an instance started inside the build. No live payment session, no real email send and no production rate limit is ever touched.
The padlock on every address we publish is checked daily, and so are the public records that stop anyone sending email in our name.
TLS · SPF · DKIM · DMARCProtocol versions, cipher suites and certificate chains per hostname. Mail records read through two independent public resolvers, so a wrong answer from one is visible rather than believed.
The system that builds and ships our software is checked as well, and every tool in it is locked to one exact version by fingerprint, so nobody can quietly swap it for something else.
Supply chainEvery action pinned to a commit hash rather than a moving tag. Least-privilege permissions declared per workflow. The pipeline definitions are themselves statically analysed, since that is where the widest permissions live.
When a safe update exists for something we use, our system writes it out overnight and asks a person to approve it. It never installs one by itself.
Fix proposalsNon-breaking upgrades only, lockfile changes only, opened as a pull request. No package install scripts run. The repository's own test suite decides whether the bump is safe. A human merges.
All of it lands in one report every morning, and a person reads it. The report names the checks that produced no answer as loudly as the ones that failed, because a check that did not run is not a check that passed.
One report, read dailyEvery repository's scan plus the account, DNS, encryption and live web sweep, assembled into a single document. Sections that did not run are counted and named rather than quietly omitted.
Finding our mistakes is worth paying for.
If you find something we missed, we would far rather hear it from you than read about it somewhere else. Tell us privately and we will fix it, credit you, and pay you when the finding is serious.
A single line that hands us a way in is worth more here than ten pages that do not. Anything that exposes customer data, or lets one person act as another, is paid in cash, and we tell you the amount when we confirm it.
- A reply from a person, within three working days. Somebody who has read what you sent. If the fix takes longer, we tell you where we have got to.
- Your name on this page. Only if you want it there. Plenty of people would rather stay anonymous.
- A free year of everything we make. Every product, whether or not the finding pays.
If it is ours, it counts
- This site and the sign-up behind it. Every page on cyberempowermenthub.co.uk, the plan builder, the checkout and the small API the site talks to.
- Anything we send you. The practice emails, the alerts, the reports, and any page they link to.
- The scam checker. What you can make it do by forwarding something to it, and what it sends back.
- Our public records. The DNS, mail and certificate records for the domains we run.
- Anything else of ours you can reach. If you found it and it is ours, send it. We would rather read the report than argue about a list.
Things we cannot fix, and things that are not a way in
- Other companies' systems. Stripe, Amazon and Cloudflare run the ground we build on. Those go to their own programmes, not to us.
- Findings with no working example. A scanner's opinion on its own is not a report. Show us what it lets somebody do.
- Settings that are not a way in. A missing header, a cipher we could tighten, a version number on show. We will still fix them, but they are not paid.
- Mail records on domains that never send. Those are set that way on purpose and checked every morning.
- People, buildings and suppliers. No phishing our staff, no ringing our providers, no turning up anywhere.
Four of them, and they are the ones you would expect
- Use your own account and your own data. Never another customer's, not even to prove the point.
- Stop as soon as you have proof. Do not read further into anything, and do not change or delete a thing.
- No denial of service. That includes load testing and any scan heavy enough to be one by accident.
- Come to us first. Send it privately and give us a fair run at fixing it before you write it up anywhere.
Stay inside those four and we will treat what you did as authorised. We will not take action against you, we will not ask anyone else to, and we will put that in writing if somebody needs to hear it from us rather than from you.
One address for all of it: a bug bounty report, a worry about your own account, or anything on this page that does not match what you are seeing. A person reads it.