Daily digest

China-Linked Fire Ant Hijacks Cisco Routers to Steal Credentials and Blind Security Logs

Fire Ant, a China-nexus espionage actor previously known for targeting VMware hypervisors, has expanded operations to compromise Cisco IOS XR routers, TACACS authentication servers, and Linux management hosts. Incident response firm Sygnia found the group using these footholds to route traffic, harvest credentials, and suppress security logging on high-value networks.

Why it matters: Compromise of TACACS and core routing infrastructure directly threatens network access-control assurance required under NIST 800-171 AC controls — review Cisco IOS XR patch levels and TACACS logging integrity as part of your next continuous monitoring cycle.

DoJ Corrects China Hacking Claim, Says U.S. Agencies Were Targets, Not Victims

The Department of Justice issued a correction to a prior press statement, clarifying that NASA, the Federal Reserve, DOE, DOJ, and other federal agencies were targeted by Chinese threat actors but were not confirmed victims of successful compromise. The original statement had implied breaches had occurred.

ValleyRAT Backdoor Hides in Signed Adware That Users Add to Antivirus Exclusions

The Silver Fox threat group is distributing the ValleyRAT backdoor disguised as QN Wallpaper, a legitimate Chinese desktop-wallpaper application, running the malware under a trusted signed process. Kaspersky reports the technique specifically exploits users who add the adware to their antivirus exclusion lists.

Why it matters: This is a reminder to audit Defender exclusion lists on Intune-managed endpoints — user-added AV exclusions are a common blind spot that undermines endpoint protection controls (NIST 800-171 SI family).

Aurora Ransomware Operators Use Cursor AI in Attacks Against 10 Targets

Threat actors linked to Aurora ransomware have been observed using the Cursor AI coding assistant to aid network intrusions, according to CloudSEK and Gambit Security. The findings stem from analysis of exposed infrastructure tied to the Russian-speaking cybercrime group across at least 10 targets.

Why it matters: If developers on your team use Cursor or similar AI coding assistants, ensure credentials and API keys used by these tools are scoped and monitored — attacker use of legitimate AI dev tools complicates detection and falls under supply-chain risk management expectations in CMMC L2.

Securing Claude Code: The New Compliance API, Local Visibility, and Identity Governance

Anthropic released Compliance API endpoints for Claude Code that give security teams visibility into agent activity — file access, shell commands, and MCP tool invocations. The article notes that activity logs alone don't establish whether an agent's access was legitimate, exposing gaps in identity governance for AI coding agents.

Why it matters: For a self-hosted AI stack, this underscores the need for identity governance around agentic AI tools acting with developer credentials — treat AI coding agents as privileged identities requiring access review, not just activity logging.

Hiding Prompt Injection in Legal Filing

A researcher documented an instance of AI prompt-injection instructions hidden within the text of a legal filing, designed to manipulate AI systems that later process the document.

Anthropic Warns Infostealer Malware Is Hijacking Claude Sessions to Drain Usage

Anthropic notified users that infostealer malware on infected PCs has been stealing active Claude login sessions, letting attackers access accounts and consume victims' AI usage quotas without needing passwords or MFA.

Why it matters: Session-token theft bypasses MFA entirely — if staff use Claude or similar AI tools with corporate accounts, enforce session-timeout policies and monitor for anomalous AI usage as part of endpoint hygiene.

Chrome Web Store Extensions Caught Stealing Crypto, Browser Data

Multiple Chrome and Edge extensions were found delivering a malware framework that deployed modules to steal cryptocurrency, sensitive browser data, and history, and to inject ClickFix social-engineering lures.

Why it matters: Review Intune browser extension allow-listing policies — unmanaged extension installs remain one of the easiest ways for malware to land on managed Windows 11 endpoints outside application control baselines.

Microsoft Asks Users to Ignore 'Antivirus Is Turned Off' Errors

Microsoft confirmed that recent Defender Antivirus updates are triggering false 'protection is off' alerts on Windows systems, and advised affected users to disregard the errors while a fix is developed.

Why it matters: Expect help-desk tickets and possible confusion with compliance monitoring dashboards that track endpoint protection status — verify actual Defender state via Intune/Defender for Endpoint console rather than relying on the local alert.

FulcrumSec Claims Manchester Airports Hack, Theft of 86 GB of Data

A threat actor calling itself FulcrumSec claims to have stolen 86 GB of data from Manchester Airports Group, including customer, booking, and travel information. BleepingComputer validated at least one traveler's record from leaked samples, revealing details beyond what MAG had initially disclosed.

The Balkanization of Virtualization Will De-Throne VMware, Which Doesn't Mind a Bit

Industry analysis argues the virtualization market is fragmenting as competitors chase AI and Kubernetes workloads rather than trying to unseat VMware in traditional VM hosting, with VMware reportedly comfortable ceding that ground while retaining its core installed base.

Why it matters: This trend validates continued investment in alternatives like Nutanix AHV, as the competitive landscape shifts toward platforms optimized for AI/container workloads rather than legacy VM parity with VMware.

Debian Votes to Let Contributors Code with AI

Debian project members voted to formally permit contributors to use AI coding assistants when submitting code, making AI-tool disclosure optional while maintaining existing quality standards for contributions.