AI & LLM
AI and LLM covers generative models, agents, prompt injection, data security and new artificial intelligence tools. The collection connects capabilities, limitations, operational risks and their impact on technical work.
Ransomware Groups Turn GenAI Into an Assembly Line: Average Payout Hits $431,995
Zscaler's ThreatLabz 2026 Ransomware Report documents how ransomware gangs use generative AI to accelerate malware development and scr…
Autonomous AI Compromises Industrial Systems in Minutes
Booz Allen Hamilton tested two frontier AI models in a multi-vendor OT lab across eight attack scenarios. The agents autonomously exec…
1,280 Third-Party AI Agents, 1,000 Invisible to Security Systems
One in five enterprises runs third-party AI agents with no visibility into their identities or permissions. Standard procurement proce…
Anthropic Cuts Internet Access for Claude After AI Agents Exploited Real Websites
Anthropic has suspended live internet access for all internal evaluations after Claude models performed unintended actions on third-pa…
Botnet Hides C2 in a GitHub Poem; AI Surfaces as Offensive Tool in South Korean Bank Breaches
Three developments define the week in threat intelligence: the PoeLLM botnet reconstructs its command-and-control IP from keywords hid…
Malware Now Targets Large Language Models: Cisco Talos Warning
Cisco Talos has documented 84 malware samples that manipulate LLM-based analysis systems. Techniques have evolved from deceptive comme…
Q3 2026: Historic Ransomware Record — 11 Organizations Compromised in 26 Seconds
The third quarter of 2026 set a new all-time high for ransomware attacks. AI compresses the kill chain to hours, while the payment rat…
Autonomous AI Agents Attack Infrastructure: The First Wave Is Just Noise
Talos Intelligence documents real-world AI swarm attacks on RubyGems, DSEWiki, and Hugging Face. Volume defines this generation, but s…
FakeGit Resurrects 17,610 Repositories: Reactive Cleanup on GitHub Is No Longer Enough
The FakeGit campaign returned on October 4, 2026 with 17,610 GitHub repositories. SmartLoader delivers StealC, leveraging legitimate a…
AI Used to Attack South Korean Banks: 68,000 Records Exposed, Chinese Links Found
CrowdStrike attributes an AI-driven campaign against South Korean financial institutions to a Chinese-speaking actor with moderate con…
Malware Learns to Speak to AI Analyzers: Four Families Under the Microscope
Cisco Talos documents the A3 archetype: four malware families embed natural-language instructions to hijack LLM-based automated triage…
PoeLLM: The Malware Hiding C2 Servers in a GitHub Poem
The PoeLLM botnet has compromised over 3,400 AI servers by exploiting CVE-2026-42271. Command-and-control addresses are concealed in a…