Google’s Risk Intelligence Group (GTIG) is warning {that a} “new and highly effective” iOS exploit equipment, dubbed Coruna by its builders has been deployed on pretend finance and crypto web sites designed to lure iPhone customers into visiting pages that may silently ship exploits. For crypto holders, the danger is blunt: GTIG’s evaluation reveals the campaigns in the end centered on harvesting seed phrases and pockets knowledge from widespread cellular apps.
Coruna targets Apple units working iOS 13.0 by iOS 17.2.1, bundling 5 full exploit chains and 23 exploits. GTIG says it recovered the equipment after monitoring its evolution throughout 2025, from early use by a buyer of a industrial surveillance firm, to “watering gap” assaults on compromised Ukrainian web sites, and at last to broad-scale distribution by way of Chinese language-language rip-off websites tied to a financially motivated actor it tracks as UNC6691.
A Crypto Lure Designed For iPhones
Within the scam-wave section, GTIG says it noticed the JavaScript framework behind Coruna deployed throughout a “very massive set” of pretend Chinese language web sites largely themed round finance. One instance cited by GTIG is a pretend WEEX-branded crypto trade web page that attempted to push guests onto an iOS system—after which a hidden iFrame can be injected to ship the exploit equipment “no matter their geolocation.”
Associated Studying
The supply mechanics matter as a result of they blur the road between conventional phishing and outright system compromise: in GTIG’s telling, merely arriving on the booby-trapped web page from a weak iPhone was sufficient to start the chain. The framework fingerprints the system to determine mannequin and iOS model, then hundreds the suitable WebKit distant code execution exploit and a pointer authentication (PAC) bypass.
GTIG tied one WebKit RCE it recovered to CVE-2024-23222, noting it was addressed by Apple in iOS 17.3 on Jan. 22, 2024.
On the finish of the chain, GTIG says Coruna drops a stager it calls PlasmaLoader (tracked as PLASMAGRID) and describes it as centered much less on traditional surveillance options and extra on stealing monetary data. In response to GTIG, the payload can decode QR codes from photographs saved on the system and scan textual content blobs for BIP39 phrase sequences, together with key phrases reminiscent of “backup phrase” and “checking account”, together with in Apple Memos, which it may well then exfiltrate.
Associated Studying
The payload can also be modular. GTIG says it may well pull down and run further modules remotely, and that lots of the recognized modules are designed to hook features and exfiltrate delicate data from widespread crypto pockets apps—amongst them MetaMask, Belief Pockets, Uniswap’s pockets, Phantom, Exodus, and TON ecosystem wallets reminiscent of Tonkeeper.
The broader arc was additionally flagged by cellular safety agency iVerify, which printed its personal findings across the identical time as GTIG’s report. “And that’s precisely what occurred once more right here, however on cellular units. Cellphone OEMs do nearly as good a job as anybody can do…”
What Crypto Customers Can Do Now
Google says Coruna “shouldn’t be efficient in opposition to the newest model of iOS,” and urges customers to replace. If updating isn’t attainable, GTIG recommends enabling Apple’s Lockdown Mode. GTIG additionally says it added the recognized web sites and domains to Google Secure Shopping to assist cut back additional publicity.
For crypto-native customers, the speedy takeaway is sensible: cellular wallets sit on the intersection of high-value belongings and high-frequency net visitors, which makes “visit-to-compromise” campaigns uniquely harmful. GTIG’s reporting suggests the rip-off funnel wasn’t nearly getting victims to attach wallets, it was about getting them onto the appropriate system, on the appropriate iOS model, so exploitation might do the remaining.
At press time, the whole crypto market cap stood at $2.45 trillion.
Featured picture created with DALL.E, chart from TradingView.com
