All news with #threat report tag

⚠ OpenAI and Anthropic models were used by attackers in a cyber-attack that targeted a municipal water and drainage utility in the Monterrey metropolitan area, Dragos reports. The incident, which unfolded between December 2025 and February 2026, involved roughly 350 artifacts, many of them AI-generated malicious scripts used as offensive tooling. According to the report, Anthropic's Claude served as the primary technical executor—handling prompt-and-response interactions, intrusion planning and deployment—while OpenAI's GPT models were used for analytical tasks and generating Spanish-language outputs. Although the OT breach was ultimately unsuccessful, Dragos warns the campaign demonstrates how commercial LLMs can accelerate and refine attacks against operational environments and recommends tighter remote access policies and stronger authentication controls.

⚠️ ThreatsDay reports a mix of blunt‑force commodity attacks and high‑impact technical flaws this week. A new MicroStealer campaign is targeting education and telecom organizations, exfiltrating browser credentials, active sessions and wallets via Discord webhooks and attacker servers. Researchers disclosed critical ICS and MOVEit vulnerabilities while analysis shows the VECT 2.0 ransomware encryptor is broken. Browsers and AI are accelerating risk vectors — patch and verify installs urgently.

🛡️ Kaspersky researchers found three malicious PyPI wheel packages — uuid32-utils, colorinal and termncolor — that covertly delivered a new malware family named ZiChatBot to Windows and Linux hosts. The packages drop platform-specific loaders (terminate.dll or terminate.so) that persist via a Registry autorun entry or a crontab and act as droppers for the main payload. ZiChatBot uses public Zulip REST APIs as its command-and-control channel, executes shellcode received from the service, and signals success by sending a heart emoji. The packages were uploaded in July 2025 and have been removed; organizations should audit dependencies, verify build environments, and monitor the published indicators.

⚠️ Cofense warns that low-skilled threat actors are increasingly abusing Vercel's v0.dev GenAI tools to generate convincing phishing pages with minimal effort. Attackers can prototype for free, purchase tokens to build pages, and use Vercel hosting—its pro tier is roughly $20/month—to deploy and tear down sites quickly. Integrations with services like Telegram, AWS, Stripe and xAI further simplify operations. Cofense advises security teams to verify sender domains, watch for urgency cues and report malicious Vercel sites for takedown.

🐧 Trend Micro researchers revealed a previously undocumented Linux implant named Quasar Linux (QLNX) that targets software developers by compromising development and DevOps environments such as npm, PyPI, GitHub, AWS, Docker, and Kubernetes. QLNX dynamically compiles rootkit and PAM backdoor modules on the host, runs fileless in memory, and employs multiple persistence methods while wiping logs and spoofing process names to remain stealthy. The toolkit includes a 58-command RAT, credential harvesting (SSH keys, cloud configs, and /etc/shadow), kernel eBPF hiding, surveillance, lateral movement, and in-memory injection; Trend Micro provided IoCs but attribution and prevalence remain unclear.

⚠️ ReversingLabs researchers describe an ongoing supply‑chain campaign called PromptMink that manipulates AI coding agents into installing malicious dependencies. Attackers publish bait packages with persuasive READMEs and LLM‑optimized documentation on registries like NPM and PyPI to increase discovery by autonomous agents and developers. The operation, attributed to North Korea’s Famous Chollima, paired legitimate‑looking SDKs with second‑layer packages carrying infostealers, later evolving to compiled Rust add‑ons, SEAs, SSH backdoors, and project exfiltration.

📩 Microsoft Defender Research disclosed a large-scale credential-theft campaign that targeted over 35,000 users at roughly 13,000 organizations using polished fake internal compliance notifications. Running April 15–16, 2026, the messages used enterprise-style HTML templates, organization-specific names and attached PDFs that redirected recipients through a Cloudflare CAPTCHA to staged authentication pages. Attackers employed an adversary-in-the-middle (AiTM) flow to harvest tokens and compromise accounts, primarily impacting US firms but seen in 26 countries. Microsoft recommends enabling passwordless authentication, using authenticator apps for MFA, turning on Safe Links and Safe Attachments, and configuring attack disruption in Microsoft Defender XDR.

🔎 A North Korea-aligned espionage group has trojanized Windows and Android clients on a regional Yanbian gaming site, according to ESET. The campaign, attributed to ScarCruft (APT37), delivered an Android port of the BirdCall backdoor (internally named zhuagou) and a trojanized mono.dll on Windows to deploy RokRAT and BirdCall. The malware harvests contacts, SMS, files, screenshots and audio, and routes command-and-control through cloud storage accounts.

🔒 Trellix disclosed on May 4 that threat actors gained unauthorized access to a portion of its source code repository and that it has notified law enforcement while working with leading forensic experts. The company, formed from the merger of McAfee Enterprise and FireEye, said it has found no evidence that its source code release or distribution process was affected or exploited. Trellix sells threat intelligence and AI-powered detection services including NDR and EDR and will share further details once the investigation concludes.

🚨This week’s telemetry shows attackers moving from quick breaches to persistent occupation across SaaS, CI/CD and hosting panels. CVE-2026-41940 in cPanel/WHM and the Linux Copy Fail bug (CVE-2026-31431) are being actively exploited alongside supply-chain compromises that weaponize developer pipelines. Social engineering — including vishing that bypasses MFA — and AI-assisted phishing kits are scaling attacks. Prioritize urgent CVEs, rotate pipeline credentials, and treat sessions and routine pipeline runs as potentially hostile.

🔐In 2025–2026, LLM-backed chat and agent systems evolved from helpful coding assistants into end-to-end development tools that materially lowered the barrier to sophisticated cyberattacks. High-profile incidents — including a 17-year-old who exfiltrated 7 million Kaikatsu Club records and adolescent and single-actor campaigns against Rakuten Mobile and multiple governments — show nontechnical actors achieving team-scale outcomes. Measured indicators worsened sharply: malicious packages surged to 454,600 and time-to-exploit collapsed to weeks. The article recommends targeting whole classes of vulnerabilities—exemplified by Chainguard Libraries—to render many supply-chain and package-distribution attacks structurally impossible.

🔵 Bluekit is a newly observed phishing kit that bundles more than 40 templates targeting services such as Outlook, Gmail, Yahoo, ProtonMail, iCloud, GitHub and Ledger. It includes an AI Assistant panel supporting models like Llama, GPT‑4.1, Claude, Gemini and DeepSeek to help draft campaign copy. Varonis found the assistant produces scaffold-like outputs that require cleanup. The platform centralizes domain purchase, phishing page setup, campaign management, granular anti-analysis controls and real-time victim session monitoring, with stolen data exfiltrated via Telegram.

🔐 Microsoft Threat Intelligence observed ~8.3 billion email-based phishing threats in Q1 2026, with volumes easing from about 2.9 billion in January to 2.6 billion in March. QR code phishing more than doubled and CAPTCHA-gated phishing surged, while link-based delivery rose to 78% and credential theft dominated payloads. Disruption of the Tycoon2FA PhaaS reduced activity but adversaries adapted; Microsoft Defender detections and mitigations are recommended.

🔒 KELA's 2026 report reveals nearly 2.9 billion compromised credentials traced worldwide in 2025, including usernames, passwords, session tokens and cookies sourced from ULP lists, breached email repositories and marketplaces. At least 347 million were obtained by infostealers operating on about 3.9 million infected machines, driven by a surge in macOS infections. The firm warns that AI-driven, autonomous attack workflows and increasing vulnerability weaponization are escalating risk for organizations.

🔒 In February 2026 researchers flagged a major shift: threat actors now deploy custom AI agents that automate attacks through the kill chain, from Active Directory mapping to rapid Domain Admin takeover. Join a technical webinar with Picus Security leaders Kevin Cole and Gursel Arici for a deep dive into Autonomous Exposure Validation. Learn how to safely ingest threat intelligence, simulate attacks, and close the gap between CTI, Red, and Blue teams to speed detection and remediation.

🛡️ Security researchers at Socket uncovered 73 additional fraudulent Open VSX extensions impersonating trusted developer tools; many now include benign code to evade scanners and later fetch a GlassWorm loader. The extensions act as thin loaders, sometimes bundling native binaries, and connect to newly created repositories to download malicious updates. Of the 73, small subsets were activated in staged waves; Socket notified the Eclipse Foundation, and most have been removed.

🔒 The AWS Customer Incident Response Team (AWS CIRT) released the March 2026 update to the Threat Technique Catalog for AWS, adding three new entries that address identity abuse, persistence, infrastructure destruction, and privilege escalation. The update highlights concrete, real-world techniques — Cognito refresh token abuse, AMI deregistration, and misuse of UpdateAssumeRolePolicy — that let attackers hide in legitimate operations. Each entry includes detection guidance and straightforward mitigations you can apply today, such as enabling refresh token rotation, protecting AMIs with Recycle Bin retention rules, and monitoring trust-policy changes.

⚠️ This week’s recap shows old techniques resurfacing alongside sophisticated new tooling that targets supply chains, enterprise remote access, and AI agents. Analysts detail fast16, a Lua-based framework predating Stuxnet that targets high-precision simulation software, and multiple active campaigns including help-desk impersonation by UNC6692 and the persistent FIRESTARTER backdoor in Cisco Firepower. Expect urgent patching, scrutiny of browser extensions and CI/CD components, and tighter monitoring of remote access and build pipelines.

🔎 Since February, Unit 42 has observed sustained operations by TGR-STA-1030 across multiple countries, with a pronounced concentration in Central and South America. The observed intrusions reuse the same tactics, techniques, and procedures previously attributed to this group, indicating continuity with prior espionage campaigns. Analysts reference The Shadow Campaigns: Uncovering Global Espionage for historical context, and advise organizations in affected regions to review detections and strengthen defensive controls.

🔒 Mandiant VP Jurgen Kutscher warns the rush to deploy AI in enterprises is reviving old cybersecurity failures as organizations neglect basic controls. During red-team engagements, Mandiant uncovered unencrypted streams, misclassified data and AI-enabled policy changes that allowed exfiltration. He urges firms to implement AI governance, revisit secure architectures and run red-team validation before uncontrolled adoption.