Information security for software delivery

ISO/IEC 27001
ISO/IEC 27001 certified
The real risk and the imagined one
The danger is loose access, not a distant desk
The picture behind the worry is an engineer in another country emailing themselves the codebase and starting a competitor. It happens rarely, and when it does the failure is almost never distance — it is that access was broad, permanent and unlogged, which is a problem an in-house hire with the same access poses just as much. The actual exposure in most engagements is duller and more common: credentials shared in a chat, a personal laptop holding a clone of the repository, a production database copied for testing and never deleted. None of those is fixed by keeping the team close; all are fixed by controls.
Approaching it as an engineering and contracting problem rather than a trust problem changes what you do. You stop trying to feel safe and start building arrangements you can point at: who owns the code, who can reach what, where the data lives, and what is revoked on the last day. The rest of this page is those arrangements. None depends on the team being in one country rather than another.
The controls
Four arrangements that keep ownership on your side of the line
These are not exotic. They are the controls any well-run employer uses; the only novelty is writing them down for a cross-border engagement rather than assuming them.
IP assigned on creation
The contract vests ownership of code in you the moment it is written, not on final payment, and covers the engineers as well as the supplier.
What it protects: There is never a window where the code belongs to someone else.
You hold the repository
The source repository, the CI pipeline and the cloud accounts are yours; engineers are granted access to them rather than working in the supplier's.
What it protects: On the last day you change nothing; they simply lose access.
Access scoped to the task
Each engineer reaches only the parts a task needs, access is logged, and it is granted with an expiry rather than forever.
What it protects: A departure or an incident touches a small, known surface.
Production data kept separate
Engineers develop against anonymised or synthetic data; real records are reached only when a task genuinely needs it, under a named approval.
What it protects: A code engagement does not quietly become a data-exposure one.
Where things live
Everything valuable sits in accounts you control
The controls above resolve to one arrangement: the assets stay in your accounts, and access to them is the thing that is granted and revoked. The engineers pass through; the code, the data and the credentials do not.
Four assets — source repository, cloud infrastructure, production data and third-party keys — all held in the client's accounts, feeding into a single access model that is scoped, logged and revocable, producing IP already owned by the client, access removable in a click and an auditable history.
Source repository
In your account, engineers as members.
Cloud and infrastructure
Your accounts, scoped roles per engineer.
Production data
Reached only under a named approval.
Third-party keys
Held in a secret store, never in code.
What we do
The access model
Scoped, logged, time-bound and revocable in one place — the single point where a departure is handled.
IP that is already yours
Assigned on creation, so nothing transfers at the end.
Access removed in a click
The last day is an account change, not a negotiation.
A history you can audit
Who reached what, and when, is recorded.
If the code, the data and the keys never leave your accounts, distance stops being the question.
Our role
We build to your controls and leave nothing on our side
- 01
Assignment written in
The contract vests IP in you on creation and binds the individual engineers, so ownership is never contingent on a final invoice being paid.
- 02
We work in your accounts
Repository, pipeline and cloud stay in your ownership with our engineers as scoped members, so there is nothing on our infrastructure to hand back.
- 03
Revocation is one action
Access is scoped, logged and time-bound, so ending an engagement or removing a person is an account change rather than a leap of faith.
Enforcing rights in court is a matter for your legal counsel, not us; our part is to arrange the ownership, the access and the data handling so the enforcement is rarely needed and the code was always yours.
Common questions
What buyers ask
Who owns the code our offshore team writes?+
You do, and the contract should say so with the ownership vesting on creation rather than on final payment — the difference matters, because payment-contingent assignment leaves a window where the code is not yet yours and a dispute could strand it. The clause also has to bind the individual engineers, not only the supplier, so that work produced by a named person is assigned through to you. Written that way, there is never a moment when the code belongs to anyone else, and the last day of an engagement transfers no rights because you already held them.
Can offshore engineers copy our source code?+
Technically, anyone with read access to a repository can clone it — that is true of an in-house engineer as much as an offshore one, and no contract makes it physically impossible. What the controls do is make it pointless and traceable: the code has no value to a competitor because ownership is assigned to you, access is scoped to what a task needs and logged so a clone leaves a record, engineers work in your accounts rather than on personal machines, and access is revoked the moment someone leaves. The protection is not a promise that copying cannot happen; it is that copying gains nothing and hides nowhere.
Should offshore engineers have access to our production data?+
As a default, no — most development is done perfectly well against anonymised or synthetic data, and a code engagement should not quietly become a data-exposure one. Real records get reached only when a task genuinely requires it, under a named approval and with the access logged and time-bound, and even then the data is not copied onto personal machines. This is the arrangement your compliance function will ask for eventually, and it is far cheaper to design at the start than to retrofit after an auditor finds a test database full of live customer records.
In this cluster
Review the controls with your team
A review of the ownership, access and data controls against your engagement, so protection is a set of arrangements you can point at rather than a hope.
Start the conversation