Skip to content

Build1 publisher2 min readPublished

Two Chinese actors hit Chrome with byte-identical shellcode before the Chromium fix shipped

Volexity says UTA0560 and JungleBamboo ran the same V8-to-kernel chain against separate targets on September 1, from separate infrastructure and with the same shellcode, while the V8 fix sat in Chromium source.

The Engineer · Build desk

Illustration accompanying Two Chinese actors hit Chrome with byte-identical shellcode before the Chromium fix shipped

What happened

  • Volexity's network monitoring service detected a spear-phishing campaign on September 1, 2026 from a Chinese actor it tracks as UTA0560, aimed at customers at multiple non-governmental organizations.
  • The emailed links abused a reflected XSS flaw on a US-based university website to bounce recipients onto attacker infrastructure hosting a chain that began with Chrome zero-day CVE-2026-85046.
  • A second Chinese actor, JungleBamboo, ran byte-for-byte identical shellcode from its own hosts, proof.gitprogram[.]com and photos.msbenefit[.]com, and delivered a different final payload.

Compiled by The EngineerSomething wrong?How this is made

Why it matters

  • exposure Keeping Chrome current bought nothing during this campaign, because Volexity says no released build contained the fix while the bug was already public in Chromium source.
  • constraint Without the Chromium commit date and the Chrome release date, nobody outside Google can size the window this attack used; the 28 days from report to exploitation is a ceiling on it.
  • contradiction The report shows two actors running the same bytes but names no supplier and no route from the upstream fix to the exploit, so treating the Chromium commit as the roadmap is unsupported by this evidence.
  • decision The last checkpoint was the link, and it resolved to a trusted university domain. Email link inspection has to follow the redirect chain.

Volexity dates one edge of the gap. A private security researcher reported the V8 type confusion to the Chromium project on August 4, 2026 [4]. Volexity's network monitoring caught it in use on September 1 [1], 28 days later [14]. The report says a fix "later entered the open-source Chromium codebase" and gives no date for that commit [15]. The interval defenders actually care about, from public fix to shipped Chrome, therefore has no measured width in this evidence. Twenty-eight days is the outer bound on it.

The chain runs three defects in series, one per boundary. The type confusion gives arbitrary read and write inside the V8 sandbox [7]. A separate WebAssembly defect, CVE-2026-87491, escapes that sandbox [8]. A Windows kernel bug, CVE-2026-85880, then escapes Chrome's sandboxed renderer and injects code into the browser process, after which the operator deploys a payload of choice [9]. On Volexity's own ordering the browser fix touches only the first step, and the first step is what the other two ride on. A released Chrome carrying the V8 patch would have broken the chain at stage one [6].

Nothing in the report says either actor found the bug by reading Chromium commits. The bug arrived from a private researcher [4], and the report does not state whether the WebAssembly defect or the kernel bug were patched upstream at all [16]. Volexity says both actors delivered the same multi-stage chain with byte-for-byte identical shellcode, with JungleBamboo hosting identical components on proof.gitprogram[.]com and photos.msbenefit[.]com and serving a different final payload [10]. Where those bytes came from, the report does not say.

For anyone defending an NGO mailbox, the reachable surface was the link. It pointed at a legitimate US-based university website carrying a reflected XSS flaw, and that redirect is what delivered the victim to cloud.shinewrist[.]net and the Files1.html landing page [2][13]. A user who checked the domain before clicking would have found a real university. The lure itself matched UTA0560 phishing that Volexity had already seen against customer organisations in March 2026, when it failed [11]. Once the exploit page loaded, the only thing visible on screen was a decoy image of a donation form built for the targeted organisation [12].

Volexity attributes the September campaign to UTA0560 and says it found JungleBamboo, which it also tracks as APT31, Violet Typhoon and TA412, exploiting the same chain against a different set of targets [3]. Two distinct activity clusters used the chain to deliver different payloads [17].

What to watch

  • A Chrome release note naming CVE-2026-85046 would let defenders measure the exposure interval instead of bounding it at 28 days.
  • Microsoft advisory detail on CVE-2026-85880 showing whether a kernel patch existed while the campaign was live.
  • Any identification of a shared exploit supplier behind the byte-identical shellcode used by both clusters.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories