Think about the filing cabinet in a busy advocate’s chamber. It holds case bundles, client letters, scanned orders and a few things that really should not be lying around. Now picture that cabinet living in the cloud: shared folders, an email system, a document-sharing tool and a handful of logins used by juniors, clerks and the occasional outside consultant. The cabinet is bigger, faster and reachable from anywhere, and so are the ways it can be left unlocked. This friendly checklist walks through what CloudSecOps means for a small legal practice or legal-information site. It is general technology guidance only and not legal advice of any kind.
Why a legal practice should care
Legal work runs on confidence. Clients hand over personal papers, and a careless setting can expose them without anybody noticing. The Ops Atlas, in its CloudSecOps entry, stresses that the cloud provider looks after the platform, while your configuration, your logins and your data remain your own responsibility. So the audit below is about the parts you control.
Part 1: who has the keys?
Grab a notebook and write out the name of everyone who can open your shared storage and email admin panel.
- Is every person and tool on that list still involved with the practice? Former interns and old consultants are the usual surprises.
- Does every login need a second step, such as a code on a phone? The Atlas describes a 2024 incident wave where affected customer accounts had no second factor and unchanged passwords.
- Does one shared login exist that several people use? If so, give each person their own and retire the shared one.
- Who can add new users or change permissions? Keep that group small.
Part 2: is anything open to the public?
The classic cloud mistake is a folder made public for convenience and forgotten. Ask your technical helper to check that no storage area is reachable without a login, and keep a rule that public access is blocked by default. The Atlas notes that a protection applied only on one folder can be removed by whoever manages that folder, so the block is more reliable when applied higher up. If a document must be shared with a client, use an expiring link rather than opening a whole folder.
Part 3: what gets recorded?
When something looks odd, you need records. Check that your cloud account keeps a log of who did what, that the log is copied somewhere ordinary users cannot edit, and that it is kept long enough to be useful. Do not assume defaults are generous. The Atlas lists default activity histories of about ninety days on large platforms and shorter periods for some sign-in records. It also recounts a 2023 review of an email-service intrusion where the affected agency had no way to look at anything older than its recording window.
Part 4: what happens to the encryption keys?
Many cloud tools encrypt stored files using keys. Ask who can use those keys and who can schedule their deletion, and find out how long a deleted key waits before it is permanently gone. A frequent misunderstanding is that regularly changing keys protects older files; per AWS documentation quoted by the Atlas, a new key leaves material sealed under the old one just as exposed. Access control and logging matter more.
Part 5: are problems owned and closed?
Many practices run a scanner or receive a vendor report and then do nothing with it. Adopt two rules from the Atlas. First, every important finding gets a named owner within one working day. Second, a finding counts as closed only when a new scan confirms it, not because someone replied “done”. Checking directly rather than trusting hearsay is a habit this blog already promotes in its explainer on following a pending matter through the official web services, and the same instinct applies here.
Part 6: exceptions with expiry dates
Sometimes you must keep an old arrangement, such as a legacy tool needing wide access. Log it in a register: why it exists, who answers for it and when it lapses. The Atlas sets ninety days as the ceiling. Review the register weekly, and treat any exception that is renewed repeatedly without progress as a warning sign.
Part 7: a regular fire drill
Once a quarter if you can, rehearse a simple scenario. A clerk’s phone is lost and the login might be compromised. Who is told first? How do you stop access, and can the logs show what was opened? Note that cutting off a login is trickier than it appears: the Atlas says temporary credentials can last many hours, so closing the way back in comes before clearing the sessions.
Part 8: keep the paperwork tidy
A short folder of dated evidence makes every later conversation easier, whether with a technical vendor, a client who asks how their papers are protected or a partner who wants reassurance. Keep the log settings, the list of people with access, the latest scan summary and the exception register in one place. Update it when something changes, and note the version of any published checklist you compared against, since baselines get revised and a reference to an old edition can mislead.
A simple measure also helps: of all the accounts the practice uses, how many are actually covered by your protective settings? If the answer is nine out of ten, the useful insight is the one account outside the net. Report the gap, not the comfortable average.
Scoring yourself
Count how many parts you can answer with confidence. Five or more suggests a healthy base. Three or four means a month of tidy-up. Fewer than three suggests starting with logging and the list of who holds keys. You do not need an expensive product to begin; a spreadsheet and fifteen minutes each week will do.
A small practice with a couple of accounts and one person looking after IT does not need a formal function. The Atlas suggests combining platform and security duties at that size. Even so, the habit of asking who can open what, and whether anyone would notice, is worth keeping. When choosing outside help, a practical article on selecting the right rental for a journey is oddly relevant: read the terms and ask what is included before you sign.