India's CERT-In warns crypto exchanges about AI-assisted attacks
India warns financial firms about AI-assisted attacks
India's Computer Emergency Response Team, CERT-In, has warned banks, financial firms and crypto businesses about a rise in AI-assisted attacks. The warning is in advisory CIAD-2026-0058, dated Nov. 4, 2026.
The advisory updates an earlier CERT-In warning from Oct. 9 about unauthorised fund transfers through payment and disbursement APIs. An API is a link that lets computer applications request a payment or transfer.
Crypto exchanges, which are platforms where people trade digital assets, and virtual digital asset service providers are listed among the affected types of business. CERT-In says wallet keys were stolen and funds were taken from crypto wallets. A wallet key is a secret code that controls access to funds in a wallet.
Key points from the Nov. 4 advisory
- CERT-In named the actor behind the campaign CRAFTY FALCON.
- The campaign abused financial firms' own cloud services and payment interfaces to move funds.
- Some firms were targeted again after earlier warnings and clean-up work that was not complete.
- Attacks were timed for holidays, weekends and other times with less monitoring.
- The public advisory names no affected exchange and gives no total loss figure.
How CERT-In says attackers reached systems
CERT-In says exposed testing and staging systems gave attackers an entry point. From there, attackers reached production systems through shared passwords, tokens or network links. Tokens are digital codes used to prove access is allowed.
Modified application packages were also used to start unauthorised API calls. The agency says cloud tools, including AWS Systems Manager and Lambda, were abused to run commands and place malicious files.
Attackers also used shared third-party servers that firms had put on allow-lists. An allow-list is a list of addresses allowed to connect. CERT-In says attackers hid in traffic from those shared servers.
The agency describes small test payments, such as one rupee, to check if an unauthorised transfer worked before larger transfers. Merchant settlement APIs, used by payment providers to settle payments to sellers, were also abused.
CERT-In says its analysis linked the campaign to AI assistance. It named publicly available offensive-security tools including ARTEX, PentAGI, Cairn and CyberStrikeAI. It also said tailored scripts and payloads were made quickly and changed to fit target responses. This part is the agency's assessment of the activity.
Protections CERT-In recommends
CERT-In calls for round-the-clock monitoring, with extra cover at weekends, holidays and other low-monitoring times. Firms that had an incident should verify through forensic checks that the intrusion was fully removed, and keep threat hunting, meaning ongoing searches for hidden attackers.
The agency recommends separating testing systems from live systems. It also recommends reducing cloud privileges, patching exposed applications quickly, and checking transaction rules on the server, not only in the application.
For API access, CERT-In recommends checking who made the request. Options it lists include mutual TLS, signed requests, and scoped API keys on dedicated vendor endpoints. Scoped keys limit what a key can do.
Monitoring should link small test payments to later large transfers. CERT-In also says firms should not rely only on blocking known indicators of compromise.
Extra protections for crypto wallet keys
For crypto service providers, CERT-In recommends keeping wallet keys in hardware security modules or cold custody. Cold custody means keeping keys offline, away from internet-connected systems.
It also recommends multi-signature withdrawal approvals, so more than one approval is needed to move funds. Firms should watch withdrawal patterns for unusual activity. Keys should be changed at once if compromise is suspected.
What is still not public
The public advisory does not name any affected crypto exchange. It does not quantify losses. CERT-In says it shared full indicators of compromise, infrastructure details and detection rules with banking, financial services and insurance firms through restricted advisories, so those details are not in the public warning.
Sources
Newisty Editorial Team covers technology, cryptocurrency, digital products, online platforms, developer tools and the wider digital economy. Our content is researched from official sources, company announcements, public documentation, market data and other primary or reputable sources. Articles are reviewed and edited before publication for clarity, accuracy and useful context.
Related Posts
Trezor and BitBox warn users of phishing emails mimicking security alerts
