BMC Daily Cyber News

This is today’s cyber news for October 6th, 2025. You can also subscribe to the newsletter and view the archive of previous headlines at daily cyber news dot com.

Attackers are delivering malicious calendar files to compromise Zimbra Collaboration Suite mailboxes, turning a routine meeting request into a takeover. The trick abuses how calendar invites are parsed, so embedded script content can run when the server or client processes event details. Victims report unauthorized mailbox rules, data theft, and persistence that starts with a simple .ics file and ends with a quiet foothold. The lure works because many mail servers auto-ingest calendar metadata to help users, and that convenience becomes an attack path. Early reporting points to a newly assigned flaw affecting recent builds, with exploitation seen in targeted phishing against self-hosted Zimbra. Parsing often occurs before traditional attachment checks, so detections can lag. Admin logs tend to show odd calendar fields and unexpected “From” variations at the moment of compromise, and we should expect copycats now that payload formats are circulating. Recommendation: Patch to the newest fixed Zimbra release and temporarily restrict automatic calendar ingestion from untrusted senders.

The U.S. Cybersecurity and Infrastructure Security Agency expanded its Known Exploited Vulnerabilities catalog with bugs already used in real attacks, spanning consumer gateways, enterprise platforms, build tools, and the ubiquitous Bash shell. Inclusion in that list is significant: it triggers federal remediation deadlines and sends a clear signal to everyone else about patch priority. These entries usually arrive with working exploits or detailed techniques that make them attractive for fast scanning. Mixed fleets and third-party hosting are especially at risk, where asset inventories are incomplete and appliances update outside normal patch cycles. Adversaries pivot quickly to newly listed items, chasing the slow patchers. Visibility gaps often sit around shadow IT, remote sites, and edge devices that lack centralized management. Teams should expect fresh scanning tied directly to the latest additions. Recommendation: Cross-check assets against the new entries and prioritize patching or compensating controls with tracked exceptions and deadlines.

Discord disclosed that a third-party customer service provider was compromised, exposing data for users who engaged with Support or Trust and Safety, while the platform itself remained unaffected. Stolen records included names, emails, usernames, limited billing details like last four digits, and a smaller set of scanned government IDs tied to age verification. Attackers reportedly attempted extortion, which is common when vendors hold identity proofs and ticket histories. The incident illustrates how support workflows concentrate high-value personal data and conversation context that can supercharge phishing. Discord revoked access, notified affected users, and pulled in law enforcement while reviewing vendor safeguards. The near-term risk is tightly targeted social engineering that references genuine tickets to build credibility. Companies that outsource support face similar exposure if access scopes are broad and retention windows are long. Recommendation: Treat unsolicited “support” outreach with skepticism and tighten vendor data retention, access scopes, and monitoring around identity documents.

Academic researchers published “WireTap,” a technique that can exfiltrate secrets from Intel Software Guard Extensions enclaves by leaning on microarchitectural and bus-level side channels rather than breaking crypto directly. SGX is meant to keep code and data safe even from the host operating system, but a long line of side-channel work shows how isolation can leak under certain conditions. The team demonstrated practical signals that can recover sensitive material, highlighting that guarantees depend on fragile hardware assumptions. Cloud confidential computing, where multiple tenants share hardware, is the most sensitive to this kind of research. Expect vendor guidance, microcode notes, or documentation updates as the findings ripple through. For now, organizations relying on enclaves should revisit their threat models and deployment patterns. Recommendation: Apply the latest mitigations and avoid colocating high-sensitivity enclave workloads with untrusted tenants.

A newly disclosed weakness in parts of the Unity ecosystem could let crafted packages or assets trigger code execution on developer machines, which then threatens the integrity of games, XR apps, or any downstream software built in those toolchains. Because Unity projects often pull plugins and assets through CI/CD pipelines, marketplaces, and contractors, a poisoned package can flow straight into builds and updates. Early analysis points to input validation and deserialization trouble during imports or editor processes. Studios with sprawling plugin catalogs and lax signing or provenance checks are the easiest targets, and community-forum assets increase exposure. We should expect proof-of-concepts that probe unsafe import paths to arrive in research repos. Recommendation: Patch quickly, pin and verify package sources, and require signed, checksummed assets in builds.

Microsoft plans to block inline SVG images by default in Outlook for the web and Windows to reduce phishing and cross-client attack surfaces, since SVG can carry active content that behaves more like a document than a static image. Email clients sanitize embedded elements differently, and Outlook’s change narrows ambiguity by refusing inline display unless policy allows. The move follows recurring reports of SVG-based lures that capture credentials or abuse browser plugins when opened from desktop clients. Marketing templates that depend on SVG may render differently depending on the mix of clients. Administrators should look for updated policy controls and rollout details as the change propagates. Recommendation: Audit templates and signatures for SVG reliance and switch to safer formats or hosted images.

Executives at organizations running Oracle E-Business Suite are reporting ransom emails that claim data theft and threaten exposure unless payments reach into the millions, with messages that borrow the tone and tactics we’ve seen from Cl0p-adjacent crews. Oracle has acknowledged the campaign and urged fast patching of known issues while investigations play out, and Google’s security team previously flagged the wave as unusually high volume. Right now these claims look uneven, likely mixing tenant-side exposure or older material to boost credibility, but the pressure is real and coordinated. Expect follow-ups that reference internal terminology or vendor ticket numbers to make the pitch sound authentic. The prudent move is to validate every claim through your own forensics and direct vendor channels rather than replying to extortion inboxes. Recommendation: Confirm posture against recent E-Business Suite updates and treat any ransom note as a trigger for independent triage.

ESET detailed two Android spyware families that masquerade as secure chat apps—one posing as a “Signal Encryption Plugin” and another as a “Pro” version of ToTok—in campaigns aimed at users in the United Arab Emirates and nearby regions. After installation, the apps rename themselves to mimic Google Play Services, then exfiltrate messages, contacts, and files while resisting removal through permission abuse. Distribution leans on third-party stores and phishing pages rather than official marketplaces, which keeps the lure active even as samples get flagged. The goal is broad handset-level surveillance that primes follow-on account takeover across personal and work apps. If you manage bring-your-own-device environments, this is exactly the kind of sideloaded threat that slips through policy cracks. Recommendation: Educate users to avoid off-store APKs and enforce mobile management policies that block sideloading on managed devices.

A persistent espionage cluster called Cavalry Werewolf is running campaigns against Russian government-aligned targets using two custom malware families. The initial lure is often a convincing local document like a procurement notice or travel memo, and once opened it deploys a lightweight foothold called FoalShell. From there, a second tool named StallionRAT handles collection, tasking, and longer-term control. The group’s operations show careful staging, including signed loaders and aged domains to blend in, plus time-based beacons that try to dodge analyst sandboxes. The victim focus looks bureaucratic and logistical rather than purely military, which hints at collection priorities around administration. The combination of selective exfiltration and decent operational security suggests they’re aiming for long dwell times. Recommendation: Harden email and document workflows with sandbox detonation, block macros by default, and hunt for the families’ distinctive command-and-control timing.

A financially motivated group operating in Chinese-language ecosystems is running end-to-end campaigns that start with search-engine poisoning and malvertising, then funnel victims through slick fake download portals. The installers look polished and name-check popular software, but they drop modular stealer payloads that harvest browser tokens, cloud console profiles, clipboards, and crypto wallets. The infrastructure overlaps with ad fraud networks and affiliate cash-out schemes, which blurs the line between gray-market promotion and outright theft. To stay ahead of blocklists, they refresh domains frequently and use content delivery fronting to hide origin servers. The victims range from consumers to small businesses that go hunting for “free” utilities and click the first result. Cleanup is messy because session tokens and OAuth grants linger even after passwords change. Recommendation: Block search ads for software in corporate browsers, allowlist official download domains, and watch for sudden token refreshes or unfamiliar grants.