Socket found 19 Chrome and Edge extensions carrying crypto-draining code, five of them bought from their original owners. The malware landed in updates after each listing had earned real installs, which is exactly what a one-time vetting pass never re-checks.
Reality
- Evidence60
- Adoption35
- Hype gap+10
- Incentives
- Insufficient
- Confidence60
Socket found that Twitch Enhanced Viewer | JeetBot attaches viewers' live OAuth tokens to playlist requests routed through its operator's proxies, exempting ten Russian-language channels. The Firefox fix has shipped and the Chrome build is in review.
Perspective Coverage
4 publishers
- Builder
- Builder 26%
- Operator
- Operator 66%
- Investor
- Investor 8%
Reality
- Evidence78
- Adoption30
- Hype gap+15
- Incentives40
- Confidence72
A dev.to writeup traces roughly 3,800 exfiltrated GitHub repositories to one trojanised Nx Console install from the official Marketplace, where the sandbox that would have contained it has been an open feature request since 2018.
Reality
- Evidence42
- Adoption28
- Hype gap+18
- Incentives62
- Confidence45
DHSeaDev's write-up on the browser card game Prismwar shows a replay check passing ten of ten matches because the recorder and the playback player both skipped the AI's defence step, so the round trip compared a bug against a copy of itself.
Reality
- Evidence45
- Adoption12
- Hype gap+10
- Incentives55
- Confidence45
Socket's threat researchers found that Twitch Enhanced Viewer sends each viewer's account-scoped OAuth token to a Russian bot vendor's proxies as a query parameter, on a redirect the extension's own code shows works without it.
Reality
- Evidence74
- Adoption52
- Hype gap+14
- Incentives62
- Confidence66
Socket traced six Chrome and Firefox listings to one publisher operation whose collector reaches for Firebase access tokens and Axiom bundle keys. Chrome pulled four in July 2026. The Firefox one is still up.
Reality
- Evidence72
- Adoption34
- Hype gap+14
- Incentives60
- Confidence63
A dev.to walkthrough counts roughly 303,800 extensions in the Chrome Web Store and 49,518 new ones in a single quarter. Agents can draft an entire listing in an afternoon, but the store still checks it against your code.
Reality
- Evidence38
- Adoption30
- Hype gap+30
- Incentives80
- Confidence40
Netskope Threat Labs puts the update between 20 and 31 July, and describes a Chrome sidebar that still earned an affiliate commission as users removed it. That says more about the update channel than about the install.
Reality
- Evidence44
- Adoption52
- Hype gap+12
- Incentives55
- Confidence46
Socket says the operators bought their way onto tens of thousands of machines and delivered the payload in a routine update. Cleanup means rotating passwords and moving crypto to fresh wallets, user by user.
Reality
- Evidence52
- Adoption34
- Hype gap+12
- Incentives62
- Confidence50
An extension allowlist pins an ID, and the ID survives a sale. The operators behind Superior bought or seeded trust that users had already granted, then spent it in an automatic update that stripped page CSP on the way through.
Reality
- Evidence55
- Adoption40
- Hype gap+12
- Incentives55
- Confidence50
One developer's Chrome-to-Firefox port of a small countdown extension came down to four keys in manifest.json. The linter that enforces them runs long before you ever see a listing form.
Reality
- Evidence62
- Adoption22
- Hype gap+12
- Incentives32
- Confidence55
Socket says the sharpest move in this campaign was buying utilities that already had users, then shipping the payload as version two. Install-time review had already run and passed, and the 80,000-user surface came with the purchase.
Reality
- Evidence62
- Adoption54
- Hype gap+14
- Incentives71
- Confidence58
A year of Sophos managed detection cases shows the AI threat that actually landed was a lookalike download page. Claude appeared in 26 of 38 incidents.
Reality
- Evidence62
- Adoption58
- Hype gap−12
- Incentives58
- Confidence55
Sophos X-Ops confirmed 38 AI-related MDR cases across twelve months. Impersonation of AI software accounted for 30 of them, and conventional payload controls were what stopped them.
Reality
- Evidence62
- Adoption54
- Hype gap−12
- Incentives68
- Confidence57