India's CERT-In Warns of Escalating AI-Driven Hacking Campaign Targeting Fintech Firms

An AI-driven hacking campaign targeting Indian fintech firms has escalated over the past few days, with non-banking financial companies (NBFCs), digital lending platforms, payment aggregators, and cryptocurrency exchanges among those affected, one of India’s key cybersecurity agencies has warned. The Indian Computer Emergency Response Team (CERT-In) on Wednesday, November 4, issued a new advisory on the escalating cyber attacks against entities in the BFSI sector and said that the “threat is active and ongoing.” The malicious actor behind the hacking campaign has been designated as ‘Crafty Falcon’ by the agency’s Computer Security Incident Response Team for Finance (CSIRT-Fin). The threat actor was found to have used AI tools to survey targets and attack them by invoking the victim’s own payment and disbursement APIs, all at machine speed. These findings were based on CSIRT-Fin’s analysis of incidents reported by the affected BFSI entities. It comes days after CERT-In became one of the few Indian government agencies to receive access to Anthropic’s Claude Mythos , a highly advanced AI model considered to be extremely capable of finding and patching security vulnerabilities. CERT-In is using Mythos to conduct small, controlled experiments on identifying software vulnerabilities while ensuring that it does not expose large volumes of potentially sensitive government and critical infrastructure data to the model hosted on third-party servers, The Indian Express previously reported. Last month, the agency also disclosed that under-testing, autonomous AI agents flouted their initial task authorisations, issued commands, invoked tools, reached the internet, and eventually gained unauthorised access to the live systems of several Indian organisations. CERT-In’s advisory on Wednesday builds on a previous warning issued on October 9, when it disclosed observing multiple cyber attack campaigns targeting fintech firms. That advisory laid out how threat actors were exploiting security flaws in internet-facing applications, exposed APIs, privileged endpoints, stolen credentials, and more to bypass controls and carry out fraudulent transactions with dummy reference identifiers to attacker-controlled mule accounts across multiple banks. Following the advisory, the affected fintech entities took steps to patch the vulnerabilities. However, they have since been “deliberately re-targeted” due to incomplete remediation, as per CERT-In. The attackers also seem to have timed the hacking campaign with the festive period to take advantage of low-monitoring windows and delayed detection and response. “All BFSI entities are strongly advised to treat this campaign as an immediate and continuing threat,” CERT-In said. Based on its analysis, CERT-In said that the attackers are using the following publicly available, AI-driven offensive security tools: -ARTEX: An open-source, multi-agent AI penetration testing and command-and-control (C2) platform. -PentAGI: An autonomous multi-AI agent penetration testing system. -Cairn: An attack search engine, in addition to the ‘CyberStrikeAI’ framework. “These tools execute complete attack chains autonomously at machine speed, generate and refine large numbers of tailored attack scripts in rapid succession and adapt their payloads based on target responses,” CERT-In said. “The supporting attack infrastructure rotates IP addresses within hours to days; infrastructure fingerprints and detailed detection rules have been shared with entities through the restricted advisories,” it added. The agency pointed out that a large number of scripts were created on the fly, signalling that they were written using AI. The threat actors targeted AWS and Lambda cloud infrastructure used by the affected organisations to execute commands and deploy malicious files. Non-production environments such as testing sites were used as entry points, with the hackers moving laterally to live production environments using shared credentials, tokens. or network paths, as per the advisory. The actors were also found to have reverse-engineered application codebases to develop tailored security exploits and initiate unauthorised API calls using modified application packages. This allowed them to bypass controls that assumed only the legitimate client application would call the APIs. Similarly, the hackers made it seem that their IP addresses were legitimate by routing their requests through trusted third-party services such as content delivery networks and app-store review tools. Before making large value transactions, the hackers carried out test transactions of Rs 1. They abused merchant settlement APIs of payment aggregators to move funds. Cryptocurrency exchanges and virtual digital asset service providers were targeted using stolen wallet keys. CERT-In also found that the attackers were quick to exploit newly disclosed security vulnerabilities, including recently uncovered flaws in the Node.js application and PHP, a widely used programming language, that had sent website operators around the world scrambling to secure their systems last month. – Since the campaign is active and escalating, CERT-In has recommended heightened 24×7 security monitoring, with increased staffing and lowered alert thresholds during holidays, weekends, etc. -Perform full eradication with forensic verification and continue periodic threat hunting; entities have been re-attacked after incomplete remediation, the agency warned. -Restrict cloud management and serverless execution to approved principals and accounts. -Enable centralised, real-time monitoring of cloud audit logs with alerting on high-risk events such as remote command execution, serverless function code updates and creation of roles, users or access keys. -Enforce instance metadata service protections and least-privilege roles for instances and functions; deny wildcard egress in security groups; configure serverless functions with no internet egress by default. -Isolate UAT/staging/testing environments from production – no shared credentials, tokens, API keys, databases or direct network paths and apply production-grade hardening and monitoring to internet-facing non-production environments. -Treat client applications and packages as untrusted: enforce server-side validation of every transaction and business rule; implement API request signing with nonce and timestamp; adopt app attestation services and certificate pinning to detect altered or repackaged applications. – Store cryptocurrency wallet keys in hardware security modules or cold custody; require multi-signature approvals for withdrawals; monitor withdrawal patterns for anomalies and rotate keys immediately upon any suspected compromise. -Monitor and govern LLM API usage and keys; LLM API traffic originating from servers or other non-development hosts may indicate an attack rig or planted operator console. -Establish a policy requiring written approval, defined scope and a time window for all penetration testing; the tools themselves do not validate authorisation. Earlier this week, CERT-In also issued new cybersecurity auditing guidelines for AI deployed by government agencies and private companies. These guidelines require reportable AI-related security incidents to be disclosed within six hours, alongside stricter requirements for risk assessment, monitoring, and independent verification of such systems.