Insights

AI Security

When your customer shows up on a leak site, check your own side first

When a ransomware group lists a health system on its leak site, the system's vendors get questions within hours. A short playbook lets a HealthTech vendor check its own exposure and answer customers without confirming or denying a claim it cannot verify.

November 19, 2026 · 6 min read

When your customer shows up on a leak site, check your own side first

A leak-site listing is a claim made by a criminal group, and nothing more until someone with access to the facts confirms it. On November 13, DataBreaches.net wrote about a group calling itself "Curious Worm" whose dark web leak site listed 12 incidents, more than half involving U.S. healthcare-related entities. DataBreaches reported that, as far as it could determine, none of the claimed incidents had been confirmed or denied by the named organizations. A Dark Web Intelligence post on X about the same site added a caution I agree with: brand-new groups sometimes pad their first listings with recycled or fabricated victims, so treat them as unverified claims until data samples or victim statements appear.

I am not going to name anyone on that list. That is the first rule of this playbook, and it applies to vendors as much as to writers.

The reason this matters for HealthTech companies is simple. When a health system you serve gets named, your phone rings. Your customer's security team wants to know if you were the way in. Your other customers want to know if they are next. Your sales team wants to know what to say. You need answers that are true, and right now you have very few facts.

What you know, and what you do not

Start by writing down three lists, in a shared document your response group can see.

What the claim says. Copy the listing details from a reputable tracker or a threat intelligence provider, not by browsing the leak site yourself. Record the date, the group name, and what the group says it has. Leave the analysis of any posted files to your counsel and an incident response firm, if it comes to that.

What you hold for that customer. Which of your systems store or process their data. What kind of data it is. Which integrations, accounts, API keys, or VPN paths connect your environment to theirs.

What you do not know. Whether the claim is real. Whether any listed data came from your systems. Whether the customer has confirmed anything.

That third list is the one people skip. It is also the one that keeps you from saying something you will have to take back.

Check your own side first

Before you talk to anyone outside, look at your own exposure to that customer. This is defensive work you can do in an afternoon.

  • Review access logs for every account and integration that touches the customer's environment or data over the last several weeks. Look for logins from unusual places, new API keys, and unusually large exports.
  • Rotate any shared credentials, tokens, or keys between your environment and theirs. Tell the customer you are doing it, so their team does not chase your change as an indicator.
  • Confirm who on your staff has access to that customer's data, and remove anyone who no longer needs it.
  • Pull your contract and business associate agreement. Find the incident notification clause and the clock it sets. Your counsel should read it the same day.
  • Preserve logs now. If this turns into a real investigation, you will want them, and retention windows roll over quietly.

If you find something, you are no longer handling a customer's claim. You are handling your own incident, and the notification clause you just pulled now applies to you.

What to say without confirming or denying

The hardest part is the first message. Here is the shape I would use for a customer who asks directly.

"We have seen reports that your organization was named on a leak site. We cannot verify those claims. We have reviewed access between our environment and yours and have rotated shared credentials as a precaution. So far we have found no evidence that our systems were involved. We will tell you right away if that changes. Here is the name and number of the person coordinating on our side."

Every sentence there is something you can stand behind. It does not say the customer was breached. It does not say they were not. It says what you checked, what you found, and who to call.

If you have not finished checking, say that instead: "We are reviewing access now and will report back by a specific time." Then meet that time.

For your other customers, say nothing unless they ask. If they do ask, keep it to your own controls: what you checked, what you rotated, and how you would notify them if you found anything. Do not name the listed organization in writing to anyone else. That is your customer's story to tell, and a claim from a criminal group is not a fact you get to spread.

Who speaks, and who does not

Pick one person to coordinate and one person to speak. They can be the same person in a small company. Everyone else, including sales and support, gets a two-line holding statement and the instruction to route questions to that person.

Write down who can approve changes to the statement. In a small HealthTech company, that is usually the CEO and whoever holds counsel's number. Agree on it before you need it, because the first call often comes after hours.

Also decide, in advance, what you will not do. You will not contact the threat actor. You will not download leaked files to "check." You will not post on social media about the listing. Those decisions are easier to make in a planning meeting than in the middle of a bad day.

Practice it before it happens

This is a good tabletop scenario because it starts with uncertainty instead of a clear breach. The setup: a threat intelligence feed tells you one of your health system customers appears on a new group's leak site. The customer's security lead calls your support line and asks, "Did this come through you?"

Then walk through it. Who takes the call? Where is the list of systems that touch that customer? Who can rotate the shared credentials, and how long does it take? Who reads the BAA? What does the first written message say, and who approves it?

If any of those answers is "I'm not sure," you have found the work. Write the playbook down, give it one page, and put the customer-facing holding statement at the top where it is easy to find.