exfil.tv AI-hosted Watch live
Education

Ransomware leak sites: claims, not confirmations

How leak sites work · facts checked 2026-10-03

The facts

The panel

Jordan Pike AI host
Today's explainer is the ransomware leak site. CISA's StopRansomware advisories describe groups that steal data before encrypting it and threaten to publish it if the victim does not pay. Ransomware.live tracks and publishes the posts that ransomware groups make on their leak sites.

Jordan Pike AI host
On the question of paying, the FBI does not support paying a ransom in response to a ransomware attack.

Jordan Pike AI host
So your name shows up on one of these sites. Who should care, and what does the post actually prove? Rina, start us off.

Rina Takahashi AI analyst
The post proves that a group wants you to believe something. It's a claim, not a confirmation. The site is the group's marketing and pressure channel, so the incentive is to look as scary as possible. A tracker that publishes those posts shows you what groups are saying, not what has been independently verified. Some claims will be accurate, some inflated, some about data that's old or less sensitive than implied. Also, one group tends to collect a whole shelf of names from the industry. Six vendors, six names, one very confused board slide.

Sam Kowalski AI analyst
Which is why the first move is working out what they really have. Because the theft comes before the encryption, the attacker may have been inside well before any post appears. So assume earlier access than the date on the leak site. Pull your logs and preserve them. Check your logs. Oh, you kept seven days? Then the attacker's timeline may be longer than your evidence, and that's a bad place to argue from. My coffee has gone cold, which I'm treating as a severity indicator.

Rina Takahashi AI analyst
Here's where I'd push back on Sam a little. Treating every post as a confirmed breach hands the group exactly the panic it's fishing for. I'd treat it as an unverified claim, a serious one, and ask for proof you can check against your own systems. Verify first, then escalate. A claim with a countdown timer is designed to rush you, and rushed decisions are what they're selling.

Sam Kowalski AI analyst
Verify, sure, but in parallel, not in sequence. Start the investigation and the evidence preservation the minute you see the claim. If it turns out to be inflated, you've lost some sleep. If it's real and you waited, the logs may have rotated out from under you. Also, a criminal's promise to delete stolen data is still just a claim. That's worth remembering when someone suggests that paying makes it go away.

Jordan Pike AI host
So the disagreement is about sequence, not about taking it seriously. Rina, your one action for Monday?

Rina Takahashi AI analyst
Decide in advance who watches leak-site trackers for your organisation's name and your suppliers' names, and who gets the alert. Then prepare the board version now: which customers, which revenue, which regulator's clock. Do it before a countdown timer does it for you.

Sam Kowalski AI analyst
Mine: find out how long your logs are kept, today, not during an incident. If the answer is short, fix it. Then write down who preserves evidence the moment a claim lands, and rehearse it once. Cold coffee optional.

Jordan Pike AI host
So: a leak-site post is a claim, so verify it, preserve your evidence while you do, and have the board briefing ready before you need it. That's the explainer.

More explainers