A weekly roundup of the most resonant events in cybersecurity: from billion-dollar money laundering schemes to autonomous AI-driven attacks.
Brazilian drug cartel and crypto brokers
Brazil's Federal Police dealt a devastating blow to an international drug syndicate suspected of smuggling at least 6.5 tons of cocaine. During a large-scale operation across four states, 44 searches were conducted and nine key figures were arrested. The investigation established that criminals used a complex network of shell companies and illegal financial brokers to convert fiat funds into cryptocurrency and move them abroad. The total amount of frozen assets reached 1 billion Brazilian reais (~$197 million). This operation confirms a troubling trend: digital assets are becoming an integral tool of transnational crime.
Attack on Minnesota water utilities
On July 26-27, unknown hackers disabled automated control systems at more than 30 water supply facilities in Minnesota. In the city of Braham, the water treatment plant came to a complete halt for three hours. The local IT agency (MNIT) activated state-level cyber defense protocols and is working to mitigate the aftermath alongside the FBI. Notably, the incident occurred amid warnings from U.S. authorities about Iranian hacker activity targeting Rockwell Automation and Allen-Bradley industrial controllers. Water treatment quality, according to officials, was not affected, but the attack itself demonstrates the vulnerability of critical infrastructure.
Ukrainian "crypto academy" turns out to be a fraudulent scheme
Ukraine's cyber police exposed an organized group operating under the guise of the "Ukrainian Financial Academy." The perpetrators deceived about a thousand citizens out of $1.11 million, promising guaranteed profits from crypto investments. The scheme was classic: advertising on social media, fake trading platforms with "success" charts, and blocking withdrawals under the pretext of fees. What made it particularly cynical was that the scammers deliberately targeted pensioners and severely ill individuals. One victim undergoing cancer treatment transferred about 500,000 hryvnias intended for therapy to the fraudsters. Four defendants face up to 12 years in prison.
SparkKitty malware: hunting for your gallery
Check Point experts discovered a new stealer called SparkKitty that infiltrated devices through official app stores. Instead of intercepting keyboard input, the virus requested access to photos and videos, then used an OCR engine to scan images for seed phrases, passwords, and QR codes. On iOS, the malware was found in the app 币coin, and on Android—in the SOEX messenger, which had been downloaded more than 10,000 times. All recognized information was sent to a command server. This once again confirms: storing wallet backups in digital form is a grave mistake that could cost you all your funds.
AI agent based on DeepSeek conducts autonomous attacks
Unit 42 specialists uncovered a Chinese-speaking hacker who used the DeepSeek neural network for fully autonomous cyberattacks. By controlling the AI via Telegram, the attacker forced the model to scan the internet through the FOFA search engine, find vulnerable systems, select exploits on GitHub, and apply them. The AI even demonstrated logic: after a failed attack on Langflow, it switched its target to the n8n platform. Although the autonomous actions were unsuccessful, in manual mode the hacker breached 11 targets and extracted data from three organizations. The attacker's infrastructure was exposed by his own mistake—an accidentally launched open HTTP server.
My expert opinion: The trend toward autonomous AI attacks is not the future—it is already a present reality. The security industry must radically rethink its approaches to defense, betting on behavioral analysis and preventive measures rather than reactive patches. Ignoring this threat could lead to catastrophic consequences for the entire digital economy.