TL;DR for CISOs: Ivanti disclosed two more Connect Secure zero-days that were chained with an earlier authentication bypass, and CISA has added both to its exploited-vulnerabilities catalog; Google and partners dismantled the XCPE botnet built into more than 1.3 million set-top boxes; and Microsoft detailed two intrusions that began with a fake help-desk call on Teams and ended in Active Directory certificate abuse.
- Internet-facing VPN appliances remain a repeat target. A patched device may still be a compromised device, so the patch ticket is only the first half of the job.
- Unmanaged consumer-grade hardware on your network can ship with a backdoor from the factory. Know where set-top boxes, projectors and similar devices sit relative to corporate segments.
- The Teams intrusions used no software flaw. They relied on a convincing request, a remote-support tool and permissive certificate settings. Your help-desk verification process is a security control.
Today's slate at a glance
| Story | Severity | Who is exposed | First action |
|---|---|---|---|
| Ivanti Connect Secure chained zero-days | Critical | Organizations running Connect Secure 22.8 or 25.1 builds | Patch, then hunt for compromise and rotate credentials |
| XCPE botnet takedown (Videostrong set-top boxes) | High | Owners of affected Android set-top boxes, including on business premises | Locate and isolate these devices |
| Teams help-desk impersonation leading to AD CS abuse | High | Organizations with Active Directory Certificate Services and open Teams federation | Review certificate templates and external Teams access |
Lead story: Ivanti patches two more Connect Secure zero-days used in an attack chain
CRITICAL · VPN / ZERO-DAYA widely deployed remote-access appliance was exploited before fixes existed, in a chain that did not leave a web shell behind. We are patching, then treating every exposed appliance as potentially compromised until reviewed, and rotating credentials that passed through it.
Key facts
| Item | Detail | Status |
|---|---|---|
| Vulnerabilities | CVE-2026-111357 (local authenticated privilege escalation, CVSS 8.8 per Rapid7) and CVE-2026-118527 (remote authenticated file upload to a known path, CVSS 4.3). They were chained with CVE-2026-104935, an authentication bypass Ivanti fixed on October 13 | Ivanti advisory as summarized by Rapid7 |
| Exploitation | Ivanti says exploitation occurred before disclosure and affected a limited number of customers. Volexity reports a cluster it tracks as UTA0576 used the chain for unauthenticated root access on a 22.8R2.4 appliance starting October 3, 2026 | Vendor and researcher statements |
| CISA | Both new CVEs were added to the Known Exploited Vulnerabilities catalog on October 30, 2026, per Rapid7 | Reported by Rapid7 |
| Fixed builds | 25.1.1.3 and 22.8R2.6 for the two new CVEs. Ivanti recommends moving from 22.8.x to 25.1.1.3 or later where possible | Ivanti guidance as summarized by Rapid7 |
| Detection | Volexity reports the activity evades Ivanti's Integrity Checker Tool, so a clean scan does not rule out compromise. Ivanti says few indicators of compromise exist | Researcher statements |
| Attribution | No threat actor has been named by Ivanti. UTA0576 is Volexity's tracking label only | Unattributed |
What happened
Ivanti updated its Connect Secure advisory on October 29 to add two further vulnerabilities that attackers had used together with CVE-2026-104935, the authentication bypass patched earlier in October. According to The Stack, the activity had been running since at least early October, and Volexity was first to observe it. Volexity says commands and their results travelled inside the HTTPS requests of the chain itself, with no separate command channel and no web shell or backdoor left on disk.
That design matters for defenders. With no persistent file to find, detection depends on request logs, outbound connections from the appliance and downstream use of credentials the device handled. Rapid7 reports that CISA added the two new CVEs to its catalog on October 30 and that Rapid7 has had authenticated vulnerability checks available since that date.
Evidence
- Rapid7: additional Ivanti Connect Secure zero-days exploited in the wild (November 2, 2026)
- The Stack: Ivanti patches two more CVEs amid highly unusual Connect Secure exploitation (October 29, 2026)
What this means for your team
Two points separate this from a routine patch cycle. First, the chain combines a pre-authentication bypass with two flaws that each look modest on their own score, which is how a CVSS 4.3 item ends up inside a root-level compromise. Prioritizing by individual score alone would have deprioritized it. Second, the reported evasion of the vendor's integrity tool means your confidence in a post-patch scan should be low. Plan for a compromise review, not just a version check.
Any credential that authenticated through an affected appliance since early October, including service accounts and VPN user passwords, is part of the exposure.
Action checklist
- Inventory every Connect Secure appliance, record its version and move to 25.1.1.3 or 22.8R2.6 or later as Ivanti advises.
- Review appliance and upstream firewall logs from October 3 onward for unexpected requests, and check the IOC Volexity published against outbound logs.
- Rotate VPN-integrated credentials and any secrets stored on the appliance, then review Active Directory activity tied to accounts that used it.
- Decide in advance whether a suspected appliance is rebuilt from a known-good image, and who authorizes that.
Google and partners dismantle the XCPE botnet built into set-top boxes
HIGH · DEVICE SUPPLY CHAINKey facts
| Item | Detail | Status |
|---|---|---|
| Scale | Google says the XCPE backdoor enrolled more than 1.3 million set-top boxes. TrendAI counted 1,472,554 infected devices across 206 models between June 3 and October 1, per The Hacker News | Google statement; TrendAI count as reported |
| Attribution | Google assesses with high confidence that Videostrong Technology Co. intentionally developed, deployed and operated the backdoor and botnet for financial gain | Google assessment |
| Capabilities | Runs at system level, reports to an MQTT-based command server, can disable Google Play Protect checks, and operators used screenshots, keystrokes and reverse shells | Google technical findings |
| Actions | Legal action against the botnet's domains and servers and against command domains of more than a dozen malware families, plus a ban on Videostrong and affiliates from building or selling certified Android devices | Google statement |
| Wider impact | Google assesses that taking down affiliate command domains severs access to at least 10 million additional compromised devices. This is an assessment, not a measured count | Assessed, not confirmed |
| Partners | Japan's National Police Agency, Lumen, HUMAN Security, Plume, Synthient and TrendAI | Named by Google |
What happened
Google Threat Intelligence Group published its account of the takedown on November 2, saying it and industry partners dismantled the backdoor the week before. Google says the operator hid the malicious code in a way that exploited a gap in its tests, loading it after the device's first boot. The Hacker News reports that owners cannot remove the backdoor themselves and that Synthient published a list of 182 affected model names, with the caution that a model name alone does not prove infection.
Evidence
- Google Threat Intelligence Group: Shutting the Door on the XCPE Botnet (November 2, 2026)
- The Hacker News: Google Disrupts 1.3 Million-Device Android Botnet (November 2, 2026)
What this means for your team
Few security programs track set-top boxes, projectors and streaming sticks, yet they sit in conference rooms, lobbies and employee homes with network access. A device that was compromised before it was unboxed will not appear in any patch report. The practical exposure is traffic proxied through your network and a foothold for lateral movement on flat segments.
Action checklist
- Search network inventory and DHCP logs for consumer media devices, and compare model names against Synthient's published list.
- Place unmanaged media and display devices on an isolated guest or IoT segment with no route to corporate systems.
- Add outbound filtering and alerting for unexpected MQTT and proxy-like traffic from those segments.
- Add a device-provenance question to procurement for any network-connected equipment, including from certified-device programs.
Microsoft details Teams help-desk lures that ended in Active Directory certificate abuse
HIGH · SOCIAL ENGINEERING / ACTIVE DIRECTORYKey facts
| Item | Detail | Status |
|---|---|---|
| Scope | Microsoft Security Research describes two human-operated intrusions that shared one opening pattern and diverged at the certificate layer | Microsoft report, October 30, 2026 |
| Opening | Attackers messaged and called users on Teams posing as IT staff and persuaded them to open Quick Assist and share the connection code. Microsoft says no vulnerability in Teams was exploited | Confirmed by Microsoft |
| Case 1 (ESC1) | Certificates were obtained for three privileged accounts, then eight more about 90 minutes later in a second wave. DCSync followed. One privileged account excluded from automated response was used to copy file-server data to an external SFTP destination with a renamed Rclone | Microsoft investigation |
| Case 2 (ESC8) | The attacker forced a domain controller to authenticate with PetitPotam and relayed it to AD CS web enrollment to obtain a domain controller certificate. The customer isolated the workstation and Microsoft saw no further progression | Microsoft investigation |
| Attribution | Microsoft does not name a threat actor in the report | Unattributed |
What happened
Microsoft published the analysis on October 30, and The Hacker News covered it on November 2. In the first case, Microsoft says Defender blocked more than 1,000 later attempts to reach files and sign in, yet data still left through the account that was excluded from automatic containment. The second case ended early because Defender raised ESC8 alerts and the workstation was isolated.
Evidence
- Microsoft Security Blog: From help desk impersonation to AD CS abuse (October 30, 2026)
- The Hacker News: Attackers Pose as IT Support on Microsoft Teams (November 2, 2026)
What this means for your team
The most instructive detail is the exclusion. An account left out of automatic containment for operational reasons became the path for data theft, which turns a convenience setting into an incident decision. The second lesson is that certificate templates and web enrollment endpoints are identity infrastructure, and many environments have never reviewed them since they were set up.
Action checklist
- Write down how your help desk proves identity, and train staff that a Teams call is not proof of anything. Require call-back through a known channel.
- Limit Teams external access to named business domains and restrict remote-support tools to approved products.
- Audit certificate templates for requester-supplied subject names and broad enrollment, and require HTTPS with Extended Protection on enrollment web endpoints, following Microsoft's KB5005413 guidance.
- List every account excluded from automated response and assign a manual compensating action for each.
Also notable
- Revolve disclosed to the SEC that attackers used automated means to obtain names, contact details, payment card numbers and expiration dates from customers with saved cards. The company told Texas that 127,200 residents were affected and Vermont that 1,328 were; it has not stated a total. The Record, October 30, 2026. SecurityWeek's November 2 headline cites at least 180,000 people; we could not open that article, so we do not repeat the figure as confirmed.
- Langflow: four CVEs in an IBM bulletin, two rated 9.8 for unauthenticated access or code execution, affect releases 1.0.0 through 1.12.8; the fix is 1.13.0, and a public Metasploit module exists, per The Hacker News. Not in the CISA catalog as of October 30. The Hacker News, November 2, 2026.
Frequently asked questions
What is the Ivanti Connect Secure attack chain?
Attackers chained CVE-2026-104935, an authentication bypass, with CVE-2026-111357 and CVE-2026-118527 to gain unauthenticated root access on Connect Secure appliances. Ivanti says a limited number of customers were affected before disclosure.
Is patching enough for Ivanti Connect Secure?
Not by itself. Volexity reports the activity evades Ivanti's Integrity Checker Tool and leaves no web shell, so organizations should review logs and rotate credentials after patching.
What is the XCPE botnet?
XCPE is a backdoor that Google says enrolled more than 1.3 million Android set-top boxes into a botnet. Google attributes it, with high confidence, to Videostrong Technology Co.
Did the Teams intrusions exploit a Microsoft vulnerability?
No. Microsoft says the activity did not exploit a flaw in Teams. It relied on social engineering, Quick Assist and insecure or overly permissive Active Directory Certificate Services settings.
What are ESC1 and ESC8?
They are certificate-services attack paths. ESC1 abuses a template that lets a low-privileged requester supply identity details, and ESC8 relays NTLM authentication to a web enrollment endpoint that lacks relay protection.
Where can I follow more breach analysis?
The CISO Platform Breach Intelligence hub carries each daily edition, and the weekly newsletter summarizes the week.
Curated by Pritha Aash, Community Head, CISO Platform
Subscribe to the weekly newsletter
More from the Breach Intelligence hub, including the BeyondTrust Remote Support zero-day, PHP2Shell technical analysis and Breach Watch Weekly: vendor access was the way in, plus vulnerability management coverage.
Corrections and takedown requests: CISO Platform is committed to accuracy and fairness. If any detail in this briefing is inaccurate, or if you represent an affected organization and would like a correction or removal, please contact us at pritha.aash@cisoplatform.com and we will review your request promptly.

Comments