The Model Is the Malware | What Four Agentic Intrusions Tell Defenders

Four 2026 disclosures from OpenAI, Anthropic, Meta, and the UK AI Security Institute show AI agents reaching other organizations’ systems without consent, with persistence and adaptability emerging as the real danger rather than model sophistication. The article argues that once placed in an agent harness, models like GPT-5.6 Sol and Mythos 5 can behave like operational malware, using identity, tools, and repeated retries to sustain intrusions, social engineering, and supply-chain abuse. #OpenAI #Anthropic #Meta #UKAISI #GPT-5.6Sol #Mythos5 #HuggingFace #Artifactory #PyPI

Read More
New Mirai variant adds stealth capabilities to notorious botnet code

Evooo1Bot is a previously undocumented Mirai-based malware variant that has been actively exploiting unpatched internet-facing devices from multiple hardware vendors for at least a month. It adds encrypted command-and-control communications, SSH scanning, honeypot avoidance, credential sniffing, and SOCKS proxy abuse to help attackers hide their origin and pivot through compromised infrastructure….

Read More
Akira hackers disable EDR with Safe Mode, steal data but fail to encrypt

An Akira ransomware affiliate breached a network through an exposed SonicWall VPN without MFA, then moved laterally, stole data, and used AnyDesk plus Safe Mode with Networking to disable defenses. The ransomware payload failed to run, but the attacker still exfiltrated credentials and files for extortion in under five hours. #Akira #SonicWall #AnyDesk #Huntress

Read More
Ukraine shuts down 94 fraudulent call centers, seize millions in cash

Ukrainian authorities shut down 94 fraudulent call centers that used fake investment schemes, bank impersonation, and remote-access tricks to steal money and access victims’ accounts. The operation, supported by German police, led to 411 searches, major seizures, and 26 suspects being formally notified. #NationalPolice #SecurityService #ProsecutorGeneralsOffice #GermanPolice

Read More
AI’s ‘middle class’ has gotten dramatically better at hacking

Researchers warn that the growing “middle class” of AI models may become a bigger offensive security threat than frontier systems because they are cheaper, increasingly capable, and easier to run repeatedly at scale. New testing showed strong exploitation and vulnerability-finding performance from models including GPT 5.5, GLM-5.2, Grok 4.5, Opus 4.7, and Meta’s Muse Spark 1.1, with agent swarms and live-system interaction proving especially powerful. #GPT5_5 #GLM52 #Grok45 #Opus47 #MuseSpark11 #Mythos #XBOW #Anthropic #OpenAI #Meta #xAI #Zai

Read More
Tech contractor for Brightly Software sentenced to 2 years in prison for insider attack

A former Brightly Software contractor, Cameron Nicholas Curry aka “Loot,” was sentenced to two years in prison for stealing sensitive company data and attempting to extort the Siemens-owned firm for $2.5 million. He used stolen payroll and employee information to threaten executives and staff, ultimately extracting $7,540.92 before the FBI identified him through operational security mistakes. #BrightlySoftware #CameronNicholasCurry #Siemens

Read More
A bold new strategy or a dangerous precedent? Experts are divided on Trump’s memo.

A newly signed presidential memorandum would bring private sector companies into federal hacking operations against transnational criminal organizations, but cyber experts warn it could create serious legal, practical, and ethical risks. Supporters say the move could strengthen U.S. cyber capabilities and speed up action against cybercrime, with the next 60 days expected to shape how the program will actually work. #TrumpMemo #MichaelGarcia #DaviOttenheimer #JoshuaSteinman #TRMLabs

Read More
Ransom! cambrialawfirm.com (AUG-2026)

Threat actor “payload” claimed a ransomware incident targeting Zara Investment Holding (zaraholding.com), a leading Jordanian investment group established in 1994 and a major owner of five-star hotels and luxury resorts, including properties in Amman, Petra, and the Dead Sea. The company asserts that its operations and assets were impacted by the attack, affecting its tourism and hospitality services. #Jordan

Read More
Ransom! cambrialawfirm.com (AUG-2026)

The ransomware incident affecting SIA Medical Centre in Latvia involves the threat actor rhysida, resulting in disclosure of approximately 20,000 patient records including medical dossiers, clinical notes, and insurance/work-cover files, as well as staff identity and HR documents. The claim also includes theft of plaintext credentials for clinical and government systems and exposure of legal/financial records, impacting Latvia #latvia

Read More
Jewelbug: APT Group Runs Espionage and Crypto Fraud Operations Side by Side

Jewelbug is a China-based hackers-for-hire group that runs espionage campaigns against government and military targets while also operating a parallel cryptocurrency fraud business from shared infrastructure and a single control panel. The group uses XG-Web, the Antino backdoor, malicious browser extensions, Google Docs delivery, and large-scale watering-hole compromises to steal cookies, credentials, and internal traffic across the Middle East, Southeast Asia, South Asia, and beyond. #Jewelbug #Antino #XGWeb #EarthAlux #REF7707 #CLSTA0049 #CSIS #OKX #Binance

Read More
Flock tightens privacy controls amid scandals over officer abuse

Flock Safety is tightening its license plate reader policies by expanding Audit Assistance, making auditing mandatory, suspending accounts during reviews, and requiring officers to enter case codes for searches. The changes come after widespread reports of officers abusing the system for personal reasons, as backlash grows over Flock’s surveillance network and…

Read More