Trapdex Subscribe

Methodology

Trapdex follows campaigns that pair a human lure, such as a fake recruiter, a fake job interview or a fake helpdesk call, with something the victim runs, installs, pastes or approves. It also follows attacks that turn a trusted third-party script against every site that loads it. What we track lists every category. It has two layers:

What we track

Trapdex tracks attacks that trick developers and employees into running something malicious, each one dated, sourced and rated for confidence. In every campaign the victim runs, installs, pastes or approves something, or a third-party script that a website trusts is altered at the vendor and runs in its visitors' browsers.

The category Other is no longer used for new campaigns.

What the victim used. A workforce campaign says whether the artifact is malicious software, a legitimate tool the attacker talked the victim into using, a command the victim pasted, or an approval the victim gave. A legitimate tool is always named, never hashed, and the page says it is not malware. Hashes are recorded only for what the attacker installed afterwards. We never record phone numbers, and tenant domains and names that contain a domain are defanged like any other indicator.

Campaign status

Every campaign page shows one of three statuses. Status answers one question: is the trap still live? It is separate from the confidence level, which says how sure we are about the campaign itself. A campaign can be rated High and still have the status Status unknown.

StatusWhat it means
ActiveA public source shows the malicious package, extension or infrastructure was still live when we last checked.
Taken downA marketplace, registry, platform or reputable public source confirms the malicious item was removed. For campaigns we track by actor rather than by a single package or extension, taken down means the campaign's known infrastructure was removed. Removal from a store does not uninstall copies already on machines.
Status unknownWe can't confirm whether this has been taken down. Treat it as possibly live. No source we use shows that it is still live or that it was removed.

Each status shows the date we last checked it. We check public sources only: removal lists, registry and marketplace listings, and vendor or platform statements. We never visit or probe attacker infrastructure to test whether it is up. Status unknown says nothing about whether the attack happened; that is what the confidence level covers.

Passive collection only

What we publish automatically

Raw collected data stays private. A publish job runs after each collection round and releases only what passes these rules:

  1. Only from sources we have approved.
  2. Every indicator is defanged, on the pages and in every download.
  3. Email addresses, phone numbers, victim usernames, credentials and tokens are dropped.
  4. Items that match our internal exclusion list are dropped.
  5. Every item is marked automated and labelled "unverified", with its source and first-seen time. Automated items never name an actor and never use confidence language.
  6. Responsible disclosure. If our own collectors flag an artifact that is still live and isn't already public in a third-party source (abuse.ch, a marketplace removal list, the Mozilla blocklist or a vendor write-up), it is not published. We report it to the marketplace first. It appears here once it has been taken down, or after the disclosure hold described on the confidence page.
  7. We can pull any item at once with a denylist. It disappears on the next update.
  8. If any check fails, nothing is published and the previous dataset stays up.

Review before interpretation

Live events are leads, not conclusions. Briefs, campaign pages, actor names and disclosure reports are all reviewed by a person before publication.

Privacy

No accounts, no cookies and no tracking. The site's few scripts run entirely in your browser: they load the live events, filter lists and search indicators.