What nobody remembers publishing
Almost every breach starts with something the organisation did not know was exposed: a test environment with a public address, a subdomain pointing at a decommissioned service, an administration panel left open 'temporarily', a database reachable from the internet through a misconfiguration.
Nobody protected them because nobody had them on the list. For an attacker, by contrast, they are the first thing that shows up.
Looking from outside
The surface inventory is done from the attacker's point of view, not from the documentation:
- Public addresses assigned in each cloud and on your own infrastructure.
- Domains and subdomains, including those pointing at services that no longer exist.
- Ports and services answering from the internet.
- Certificates issued for the organisation's names, which reveal systems nobody sometimes remembers.
- Leaked credentials in public repositories or known dumps.
What is found is compared with what you believe is exposed. The difference is the most urgent work list there is (just as in the survey of what you inherited, but looking outwards).
Reduce before protecting
For each exposed item, the first question is not 'how do I protect it?', but 'does it have to be exposed?'. Often the answer is no:
- Test environments and administration panels are not published; they are reached through private access.
- Orphaned subdomains are deleted.
- What does have to be public goes through the exposed zone, with its protection, as in designing by zones.
An inventory that does not age
The surface changes every time somebody deploys something. That is why the inventory is not an annual exercise, but a periodic, automatic check that raises an alert when something newly exposed appears that was not planned.
Knowing what has to be protected enables the next step: deciding with which controls, and building on what already exists instead of inventing it.
How we apply it: Cybersecurity.