CloudSEK found a rented malware service whose fake page could send card fields as they were typed, while a two-stage Android package contained routines to intercept OTPs and set call forwarding.
A rented credential-stealing Android malware panel targeted customers of six Indian banks and card issuers with a sequence built around the way card fraud is verified.
CloudSEK found 50 tenant accounts on the service, including 34 paid licences. Nineteen approved tenants were live on November 10. Four days later, its institution-by-institution table recorded a page and Android package for three financial brands, and an Android package for the other three.
The panel gave each tenant a fake app page, a branded Android package and a separate exfiltration channel. The page collected payment and identity details. A two-stage Android payload could then gather text messages and one-time passwords. The same payload contained routines to set call forwarding.
Party Panel joined three signals banks often handle separately—the card data, the OTP and the follow-up call. Defenders should correlate changes to all three instead of treating each event as an independent alert.
CloudSEK found 50 tenant accounts, including 34 paid licences, in a panel that let operators select a target brand and page design and build a matching Android app.
CloudSEK decrypted and statically examined both Android stages without running them. The researchers say they did not access victim records. The report establishes the panel’s configuration, tenant scale, malware capabilities and dated artifacts; it provides no infected-device count or completed-fraud total.
The page collected before the app installed
The first collection step happened in the browser.
CloudSEK analyzed a fake app-store page dressed as a credit-card application from a financial institution. It copied familiar store elements and asked for payment and identity information before directing the visitor toward an Android package.
The page’s JavaScript sent entered data to the panel as the user typed. Reaching the final submit button was unnecessary for partial information to leave the browser.
Tenants could choose among lure designs, including app-store clones, online-banking pages and direct-download pages. They selected an impersonated brand, a form and a messaging destination; the panel assembled those choices into a branded package.
That boundary matters in an investigation. A visit to the lure may expose information entered into its form. Installing the package creates a broader device-compromise case. Responders need to establish which stage the user reached before deciding what must be invalidated.
Two Android stages divided delivery from collection
The Android components had different jobs.
CloudSEK’s static analysis found that the first package presented itself as a financial app while preparing a second component. Its code also attempted to weaken Android protections during installation.
The second component could hide its launcher icon and persist after a restart. Its routines could read and forward incoming text messages, collect OTPs, upload existing messages and gallery material, send messages from the device and change call-forwarding settings.
Static analysis documents capability. A lure establishes targeting. An installed payload establishes device exposure. Device, account and transaction evidence must establish what occurred next.
CloudSEK reported a fake app-store-style page leading to a two-stage Android chain: a dropper followed by a hidden SMS-stealing payload.
The verification loop was the target
Card-not-present fraud often fails when one check remains outside the operator’s reach. Payment details are incomplete without the OTP. The OTP may arrive, but a verification call can still reach the legitimate customer.
Party Panel was configured to reach across those boundaries.
The fake page could collect the payment details. The hidden payload could relay the OTP. Call forwarding could move an incoming verification call away from the customer. Stored messages and images could provide more context for impersonation or account-recovery attempts.
The useful connection in CloudSEK’s findings is the sequence. The service was built around several parts of a verification process rather than one stolen artifact.
A bank may see a valid card number, a correct OTP and an answered call as separate checks. On a compromised phone, the same malware operation can influence all three.
CloudSEK found a fake page configured to collect payment-card data and Android routines capable of intercepting OTP and SMS messages and setting call forwarding.
Fifty tenants turned one build system into many lures
The exposed panel showed how the service scaled.
Of its 50 tenant accounts, 34 held paid licences lasting from seven to 60 days. Nineteen approved tenants were live on November 10 alone. The themes covered six banks and card issuers, alongside government-service lures aimed at groups including motorists, farmers and pensioners.
CloudSEK assessed with medium confidence that the operator was based in India, citing language, time-zone formatting, hosting and target selection. That is a circumstantial geographic assessment; it does not identify an individual or establish who controlled every tenant account.
The tenant count shows a repeatable service. It does not establish 50 distinct operators or 50 successful campaigns. The stronger evidence is the combination of active licences, generated packages and artifacts that were still available after tenants began operating.
CloudSEK recommended correlating call-forwarding activation with new-device OTP use around card-not-present transactions rather than treating each event alone.
Correlate call forwarding with new-device OTP use
The best defensive signal in the report is a pair of events.
CloudSEK recommends watching for call-forwarding activation and OTP use from a new device around a card-not-present transaction. Either event can have an innocent explanation. Together, close in time, they describe the verification loop the malware was designed to control.
I would investigate that combination as a possible device-compromise case immediately.
For banks and fraud teams, the work is to connect the transaction, OTP delivery, device change and available call-routing signal before closing them as independent alerts. A customer report that verification calls stopped arriving belongs in the same case as recent OTP and card activity.
For mobile responders, establish whether the user only visited the fake page or installed the package. If installation occurred, isolate the phone, preserve its app and messaging evidence, and use trusted support to remove the malicious software and review what it could access. Contact the issuer and mobile carrier through official channels from a clean device, cancel unauthorized forwarding and replace exposed payment or account access.
Financial institutions should publish the exact official identity of their mobile app and make sideloading warnings specific. Takedown work must cover the panel, hosting and messaging infrastructure as well as individual lure pages, because the service can generate replacements.
Break the chain where the signals meet
Party Panel’s design coordinated several capabilities. A fake page was configured to send payment details as they were typed. Android routines could reach the OTP and set call forwarding, placing the bank’s human verification path inside the same fraud design.
That coordination creates a detection opportunity too.
When call forwarding changes, a new device uses the OTP and a card-not-present attempt follows, the combined signal deserves a faster decision than any one alert would receive alone.






