Skip to main content
promotional banner for can you prove your CC^ checklist
Four Days Before the Patch: Closing the Zero-Day GapVulnerability and Software Security
5 min readFor Bank Information Security Officers

Four Days Before the Patch: Closing the Zero-Day Gap

Attackers exploited the PHP2Shell flaws four days before PHP released emergency fixes on October 21. CISA added both CVEs to its Known Exploited Vulnerabilities (KEV) catalog on October 26. The attack targeted Check Point Security Gateways, then moved to VMware ESXi hosts, deploying a wiper. If your patch SLA is measured in weeks, the attacker's timeline was shorter than your change window.

This guide provides a process to reduce exposure, integrate threat intelligence into vulnerability management, and ensure you're not already compromised.

1. Understanding the Attack

The attack chain is instructive. The attacker used PHP2Shell (CVE-2026-109618 and CVE-2026-109619) for unauthenticated remote code execution through the gateways' Mobile Access portal. They then used a newly disclosed zero-day, CVE-2026-123039, to gain root access. CISA also warned of PHP2Shell being used with the Fragnesia Linux kernel flaw for similar purposes. From the gateway, the attacker installed PHP webshells and the TSGA backdoor, located ESXi hosts with the gogo scanner, and ran the destroy_v6 wiper.

For a bank, two points are critical. First, internet-facing remote access infrastructure is the entry point. Second, the target was hypervisor infrastructure, where a wiper threatens availability across many systems. A ransom note wasn't the goal; the dropped note explicitly rejected negotiation. Your recovery plan can't rely on a payment path.

2. Preparation Essentials

  • Asset inventory with exposure tagging: every internet-facing gateway, VPN, and remote access portal, plus the management plane for each.
  • A KEV feed ingest. CISA publishes the catalog as JSON and CSV; pull it on a schedule.
  • Log retention from gateways, including firewall logs, with a defined baseline so gaps are detectable.
  • Offline or immutable backups of ESXi hosts, VM configurations, and datastores, with a tested restore process.
  • Network path documentation showing which systems can initiate SSH to the hypervisor management network.
  • A named owner with authority to approve emergency changes outside the normal cycle.

3. Step-by-Step Implementation

Step 1: Tier Your Exposure. Classify assets into three groups: internet-facing with remote code execution potential (Tier 0), internal but privileged such as hypervisors and management servers (Tier 1), and everything else (Tier 2). In this campaign, the gateways with an exposed Mobile Access portal were Tier 0, and the ESXi hosts were Tier 1.

Step 2: Automate KEV Matching. Schedule a job that pulls the KEV catalog and matches it against your inventory by vendor and product. A KEV match on a Tier 0 asset should open a ticket automatically. A minimal pattern:

curl -s https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json \
  | jq '.vulnerabilities[] | select(.vendorProject | test("Check Point|PHP|VMware"; "i"))'

Adjust the filter to your inventory. The goal is for a KEV entry to reach a human the same day, not at the next monthly review.

Step 3: Set Exploitation-Driven SLAs. Replace severity-only timelines with exploitation status. A workable structure: KEV-listed on a Tier 0 asset triggers emergency change and mitigation within days, with compensating controls (disable the Mobile Access portal, restrict source IPs, or place it behind additional controls) applied immediately if the patch can't be staged. Document the thresholds in your vulnerability management standard so examiners can see the logic. This supports the risk-based patching expectations in the FFIEC IT Examination Handbook and maps to the Identify and Protect functions of the NIST Cybersecurity Framework (CSF).

Step 4: Apply the Vendor Fix. Check Point recommends installing the November Jumbo Hotfix on all Management Servers and Security Gateways, prioritizing gateways with internet-facing Mobile Access portals. It addresses both PHP2Shell and the DynamicCLI privilege escalation flaw used in these attacks. Patch Management Servers too; don't stop at the perimeter.

Step 5: Restrict Gateway-to-Hypervisor Paths. The pivot to ESXi relied on the gateway reaching hypervisor hosts. Enforce Least Privilege at the network layer: ESXi management interfaces should accept SSH only from a dedicated administrative jump host, and SSH on ESXi should be disabled when not in use. Run Segmentation Testing to confirm the gateway cannot initiate sessions to the management VLAN.

Step 6: Route Threat Intelligence into Incident Response. Pre-build the playbook for a KEV-listed flaw on a Tier 0 asset: scope exposure, hunt for Indicators of Compromise (IoCs), patch, and re-verify. Assign who decides on isolation. Treat attribution as a separate, lower-priority track.

4. Validation

Patching doesn't prove you weren't breached before the patch. For any gateway that was exposed, check Check Point's published compromise indicators:

  • Session files containing upload_progress_pwn.
  • Unexpected CPconfigdaemon processes or binaries.
  • Unexplained gaps in firewall logs.
  • Unauthorized SSH activity from the gateway to internal VMware ESXi hosts, which is a clear sign of breach.

Search your SIEM for gateway-sourced SSH to hypervisor addresses over the full retention window. Compare log volume per hour against your baseline to surface gaps. If you find any indicator, assume root-level compromise and treat the host as untrusted. Then test your backups: restore an ESXi configuration and one VM from offline media, and record the time it takes.

5. Maintenance

  • Daily: KEV feed job runs; new matches to Tier 0 assets page the on-call owner.
  • Weekly: Review open exceptions and compensating controls; any control without a patch date gets escalated.
  • Quarterly: Re-run Segmentation Testing between perimeter devices and the hypervisor management network. Test backup restoration.
  • After each incident or near miss: Measure the interval between first public exploitation signal and your mitigation. Track it as a metric to report to your risk committee.

A Note on Attribution

Check Point tentatively linked TSGA and TSProxy to PhiliKit, a framework previously seen in Ivanti appliance attacks and associated by ESET with the cluster UNC5221, but could not attribute the activity with confidence. The destructive stage carried political messages, which raises the possibility of a false flag. For your risk assessment, the practical conclusion is that attribution shouldn't change your response. Whoever is behind it, the defensive controls are identical, and delay spent debating motive is delay the attacker uses.

Promotional banner for the CUI buildout guide

You Might Also Like