← Carolina Clear Tech

Cyber Threat Brief

2026-11-16

Listen to this brief (12:31)

Download MP3
Show Notes

Show Notes - 2026-11-16

Stories Covered

CVEs Referenced

CVE-2021-44228

Indicators of Compromise

Domains: nofrillsdns[.]com, moments-notice[.]com, planetmaison[.]com, newbooker[.]com, reallifeaccounting[.]com, efxisnqldelyr5ra25oqfrsut54b34g26e7zrsiuupsoekjof2gho2id[.]onion

IP Addresses: 0.4.9.17

Read the full brief

Get tomorrow's brief in your inbox

Protect Your Business

Need a security assessment? See our cybersecurity packages.

View Services

Today: BO Team exploited Log4Shell in VMware vCenter during an intrusion that reached a software developer's code-signing infrastructure. Barracuda disclosed a cloud compromise involving an unrevoked third-party access key, while Microsoft warned that phishing emails are impersonating notifications for its resolved Microsoft 365 outage. Prioritize vCenter remediation checks, third-party credential reviews, and investigation of outage-themed email clicks.

Critical Alerts

BO Team exploits VMware vCenter Log4Shell and compromises code-signing infrastructure (CVE-2021-44228)

Summary: Kaspersky attributed a months-long intrusion at an unnamed Russian software developer to BO Team. Attackers entered through an exposed SSH service, exploited CVE-2021-44228 on a VMware vCenter 7.0.3 server for root-level command execution, and compromised infrastructure including backup servers, security management systems, and a code-signing server. Enrichment identifies Log4Shell as CISA KEV-listed and ransomware-linked, with EPSS 1.000 at the 100th percentile; its listed remediation deadline, December 24, 2021, is historical, not a new deadline.

IOCs & Detection

BO Team, Microsoft 365 phishing, CoffeeLoader, and SANSAR detection leads

Summary: The following indicators and behaviors were identified in the individual incident reports. Confidence reflects their usefulness as investigation leads, not proof that a matching system is compromised; filenames, legitimate utilities, and shared infrastructure require corroboration.

Business & Infrastructure Threats

Barracuda cloud intrusion began with an unrevoked third-party key

Summary: Barracuda says an attacker used a cloud access key previously issued to a third-party AI service provider, then deployed ARTEX to expand access across 10 cloud environments, including environments managing its XDR platform. The company found no evidence that the attacker accessed XDR customer integration credentials to pivot into customer networks, but notified some customers and advised precautionary credential reissuance. Barracuda learned of the breach on November 4, two days after the attacker's last activity, according to the report.

AI-agent campaigns target exposed application gateways and provider keys

Summary: Amazon captured agent records describing attempts to steal AI-provider keys from internet-exposed applications including LiteLLM, Dify, One API, Langflow, and RAGFlow. Amazon cannot verify the outcomes described in those records and assesses ARTEX involvement and separate China-linked operators with moderate confidence. Separately, a report on an FBI warning describes ARTEX use against organizations worldwide, including a ConnectWise supply-chain breach affecting Microsoft 365 customer environments.

SANSAR hides Linux GPU mining by falsifying monitoring data

Summary: Elastic documented a Linux cryptomining intrusion beginning through an exposed Ray job agent, followed by an unauthorized SSH key and root-level deployment of the SANSAR toolkit. Its NVML wrapper reports GPU utilization as 0% and substitutes temperature and power readings, while a kernel rootkit conceals processes, files, and sockets. Persistence includes a DKMS-built module and an initramfs loader, making routine process listings and host-reported GPU utilization unreliable on an affected system.

Windows / AD Security

Microsoft 365 outage notifications repurposed for phishing

Summary: Microsoft attributed the November 12 Exchange Online and Outlook disruption to a technical network issue, not a cyberattack, and marked incident EX1550543 resolved on November 13. Attackers reused that real identifier in fraudulent service-health emails claiming held messages would be deleted within 24 hours unless recipients clicked a recovery link. Microsoft characterized the activity as consistent with credential phishing, while stating that credential collection and malware delivery were not confirmed.

CoffeeLoader update adds a downloader and Windows persistence changes

Summary: Kaspersky detected a CoffeeLoader variant in August 2026 that uses a patched 7z.dll and an updated Armoury packer relying on OpenGL rather than OpenCL for GPU-assisted decryption. A new downloader establishes Run-key persistence, injects into Windows processes, and retrieves encrypted payloads over HTTPS. It also creates a registry setting intended to prevent users from disabling its startup entry through Task Manager.

General Security News

Italy orders Facebook gallery-consent changes and fines Meta €805 million

Summary: Italy's Garante found that Facebook's Android gallery-suggestion feature accessed photos, videos, and metadata without separate explicit consent, and fined Meta €805 million. The authority ordered separate consent and default-off behavior within 60 days, with progress reporting within 30 and 90 days; these obligations apply to Meta, not customer IT teams. Meta plans to appeal, calls the fine disproportionate, and says users grant device permission and can change it in settings.

Patch Priority

Vulnerability Disclosures

Tor 0.4.9.17 strengthens defenses during ongoing relay attacks

Summary: Tor released version 0.4.9.17 on November 13 with several security fixes, including changes to IPv6 handling in relay DoS protections. The Tor Project urged relay operators to update without delay as large volumes of client circuits overwhelmed the network. One operator described continued disruption after upgrading; their warning about possible deanonymization was an individual assessment, not a warning issued by the Tor Project.

Trends & Context

Old vulnerabilities, exposed management services, and retained credentials remain practical entry points even when attackers use AI-assisted tooling. Across these incidents, prioritize access restrictions and credential lifecycle controls alongside patching, and corroborate telemetry when attackers can manipulate the monitoring layer itself.