Weekly Recap Highlights AI Skill Malware, Record 31 Tbps DDoS, Notepad++ Supply Chain Abuse, And LLM Backdoors

Published:

Cyber threats continue to evolve beyond traditional malware and software exploits, increasingly embedding themselves into trusted platforms, AI ecosystems, developer tools, and communication services that organizations rely on daily. This past week highlighted a consistent trend where attackers leveraged trust rather than brute force, quietly slipping into legitimate workflows, marketplaces, and update mechanisms that already have access to sensitive environments. As artificial intelligence, cloud platforms, and automation tools become deeply integrated into enterprise operations, adversaries are following the same paths, turning convenience and scale into attack vectors. The result is a rapidly expanding threat surface shaped by ecosystem abuse rather than perimeter failure.

One of the most closely watched developments involved OpenClaw, which announced a partnership with Google’s VirusTotal to scan AI skills uploaded to its ClawHub registry. The move followed growing concern within the cybersecurity community over malicious skills discovered in the public marketplace, where attackers have been planting malware to target developers and organizations experimenting with agentic AI. Researchers warned that autonomous AI tools with persistent memory, broad permissions, and user driven configuration can amplify risks such as prompt injection, data leakage, and unvetted dependencies. Trend Micro confirmed active discussions on the Exploit.in forum around using OpenClaw skills for botnet activity, while Veracode reported a surge in npm and PyPI packages using the claw naming convention, jumping from almost none to more than 1,000 within weeks. These developments reinforced warnings that open source agent ecosystems require stronger user awareness and security discipline, as unsupervised deployment can turn theoretical risks into operational threats.

Large scale attacks and trusted software abuse further defined the week’s threat landscape. Cloudflare attributed a record setting 31.4 Tbps distributed denial of service attack to the AISURU or Kimwolf botnet, marking the largest known DDoS event to date despite lasting only 35 seconds. The same botnet was linked to another campaign observed in late December 2025, as overall DDoS activity surged significantly throughout the year. At the same time, a sophisticated supply chain compromise targeted Notepad++, where attackers selectively redirected updater traffic from the WinGUp update mechanism to malicious servers between June and October 2025. Although access to a third party hosting provider was partially disrupted, valid credentials allowed the attackers to continue rerouting update traffic into December. The campaign, attributed to Lotus Blossom, underscored how update mechanisms, even from well known domains, remain a high value target when verification controls are weak.

Artificial intelligence itself also emerged as both a target and an enabler of abuse. Docker patched a critical vulnerability in its Ask Gordon AI assistant after researchers demonstrated remote code execution via malicious metadata injected into container images, exploiting misplaced trust in contextual information. Microsoft disclosed the development of a new scanner designed to detect hidden backdoors in open weight large language models by identifying behavioral signatures and poisoned data patterns. Meanwhile, security researchers flagged increasing exposure of OpenClaw gateways, widespread prompt injection risks on AI driven platforms like Moltbook, and malicious npm packages using Ethereum smart contracts as command and control resolvers to evade takedowns. Additional findings ranged from fileless Linux post exploitation frameworks operating entirely in memory to the continued growth of illicit online marketplaces and large scale ecosystem abuse. Together, these incidents illustrated how attackers are blending AI, automation, and trust exploitation into coordinated strategies that challenge traditional security assumptions across the global digital landscape.

Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem. 

Related articles

spot_img