On September 22, 2026, CISA added CVE-2026-93616 to the Known Exploited Vulnerabilities (KEV) catalog, confirming what many of us in the IR community suspected the moment the advisory dropped: threat actors are actively exploiting a path traversal vulnerability in Check Point's core management infrastructure in the wild, right now.
This is not a perimeter VPN bug. This is the Security Management Server — the brain of your Check Point deployment. If an attacker lands arbitrary code execution on your management server, they don't just own a box. They own your policy, your logs, your forensics trail, and potentially the keys to every gateway in your estate. For defenders, this is a drop-everything event.
Who Is Affected
Per the KEV entry and vendor guidance, the following Check Point products are vulnerable:
- Security Management Server (SMS)
- Multi-Domain Security Management Server (MDS / Provider-1)
- Log Server
- Multi-Domain Log Server
- SmartEvent Server
These components run on Check Point's Gaia OS (Linux-based) and are frequently co-located. An MSSP or enterprise running Multi-Domain management is looking at a blast radius spanning multiple customer or business-unit environments from a single compromise.
Technical Analysis
Vulnerability Mechanics
CVE-2026-93616 is a path traversal vulnerability in a web-facing component of the Check Point management stack. The flaw allows an unauthenticated, remote attacker to escape the intended web root directory and upload arbitrary scripts to attacker-controlled locations on the filesystem, which are subsequently executed by the underlying service.
From a defender's perspective, the attack chain looks like this:
- Reconnaissance: Attacker identifies an exposed Check Point management web interface (HTTPS, typically TCP/443). Internet-exposed management portals are the primary target, but any network segment with reachability to the management interface is in scope.
- Traversal: A crafted HTTP request containing path traversal sequences (
../, URL-encoded variants such as%2e%2e%2f, double-encoded variants like%252e%252e%252f) reaches a file-handling endpoint that fails to canonicalize and validate the requested path. - Upload: The attacker writes a script payload — historically we see JSP webshells, PHP one-liners, or shell scripts — to a web-accessible or service-executed directory.
- Execution: The attacker (or the service itself, on the next request/service cycle) triggers the payload. On Gaia-based systems this commonly manifests as the web service process (Java/Tomcat or httpd workers) spawning
/bin/sh,bash,python, orperl. - Post-exploitation: Credential harvesting from Gaia's configuration store, theft of SIC (Secure Internal Communication) certificates, policy manipulation, log tampering, and pivoting to managed gateways or the internal network.
Severity Assessment
Unauthenticated + remote + arbitrary code execution + management plane = this sits at the top of the severity stack regardless of the final published CVSS score. Treat it as critical. CISA's KEV addition is the authoritative signal: this is not theoretical, and there is no "wait for the next maintenance window" option.
Regulatory Forcing Functions
CISA's KEV entry references two binding obligations for federal civilian agencies — and they are excellent guidance for everyone else:
- BOD 26-04 (Prioritizing Security Updates Based on Risk): Requires applying vendor mitigations within the mandated timeline for KEV-listed vulnerabilities. Private-sector organizations should adopt the same SLA — KEV items deserve a 72-hour-or-less remediation posture.
- CISA Forensics Triage Requirements: Before (or alongside) patching, capture forensic evidence. If your management server was internet-reachable and unpatched, you must assume potential compromise and triage accordingly. Do not simply patch and move on — patching erases nothing that already happened.
If mitigations cannot be applied, BOD 26-04 guidance directs discontinuing use of the affected product. For most organizations that translates to: isolate the management interface from all untrusted networks immediately.
Detection & Response
The detections below target the observable behaviors of this attack chain: traversal attempts in web logs, the management web service spawning unexpected child processes, and webshell/script artifacts written to web-accessible paths. Tune path constants to your deployment (Smart-1 appliances vs. Gaia on open server vs. VM).
---
title: Check Point Management Web Service Spawning Shell or Interpreter
id: 3c8f1a72-9b4e-4d21-ae67-2f5c8d901234
status: experimental
description: Detects the Check Point management web stack (Tomcat/Java or httpd) spawning a shell or script interpreter — a strong post-exploitation indicator for CVE-2026-93616 arbitrary script execution.
references:
- https://www.cisa.gov/known-exploited-vulnerabilities-catalog?search_api_fulltext=CVE-2026-93616
- https://attack.mitre.org/techniques/T1059/
author: Security Arsenal
date: 2026/09/23
tags:
- attack.execution
- attack.t1059
- attack.t1190
logsource:
category: process_creation
product: linux
detection:
selection_parent:
ParentImage|endswith:
- '/java'
- '/httpd'
- '/apache2'
selection_child:
Image|endswith:
- '/sh'
- '/bash'
- '/dash'
- '/python'
- '/python3'
- '/perl'
- '/curl'
- '/wget'
- '/nc'
- '/ncat'
condition: selection_parent and selection_child
falsepositives:
- Check Point maintenance scripts and health-check automation — baseline parent/child pairs during a quiet window and allowlist known-good hashes
level: high
---
title: Path Traversal Attempt Against Check Point Management Interface
id: 7d2e5b91-4c3a-48f6-b912-8a4e6f102738
status: experimental
description: Detects path traversal sequences and script-upload indicators in HTTP requests directed at Check Point Security Management / SmartEvent web services. Apply to reverse proxy, WAF, or web server access logs fronting the management plane.
references:
- https://www.cisa.gov/known-exploited-vulnerabilities-catalog?search_api_fulltext=CVE-2026-93616
- https://attack.mitre.org/techniques/T1190/
- https://attack.mitre.org/techniques/T1083/
author: Security Arsenal
date: 2026/09/23
tags:
- attack.initial_access
- attack.t1190
logsource:
category: webserver
product: linux
detection:
selection_traversal:
cs-uri|contains:
- '../'
- '..%2f'
- '%2e%2e'
- '..%5c'
- '%252e%252e'
- '..;/'
selection_upload:
cs-method:
- 'POST'
- 'PUT'
condition: selection_traversal and selection_upload
falsepositives:
- Rare; legitimate management portal traffic should never contain traversal sequences. Scanner traffic may match — corroborate with source reputation
level: critical
---
title: Script File Created in Check Point Web-Accessible Directory
id: 91f4c2d8-6e7b-43a5-9d01-5b8c3e247190
status: experimental
description: Detects creation of executable script files in web-served or upload directories on Check Point Gaia management servers — consistent with the arbitrary script upload phase of CVE-2026-93616 exploitation.
references:
- https://www.cisa.gov/known-exploited-vulnerabilities-catalog?search_api_fulltext=CVE-2026-93616
- https://attack.mitre.org/techniques/T1505.003/
author: Security Arsenal
date: 2026/09/23
tags:
- attack.persistence
- attack.t1505.003
logsource:
category: file_event
product: linux
detection:
selection_ext:
TargetFilename|endswith:
- '.jsp'
- '.jspx'
- '.php'
- '.sh'
- '.py'
- '.war'
selection_path:
TargetFilename|contains:
- '/web/'
- '/htdocs/'
- '/webapps/'
- '/uploads/'
- '/tmp/'
- '/var/tmp/'
condition: selection_ext and selection_path
falsepositives:
- Vendor hotfix installations and legitimate admin file transfers — correlate with approved change windows
level: high
// Hunt: CVE-2026-93616 — traversal probes and anomalous traffic to Check Point management plane
// Ingest Check Point syslog/web access logs via CEF into CommonSecurityLog
let lookback = 14d;
let mgmt_ips = dynamic(["10.0.0.15", "10.0.0.16"]); // <-- replace with your SMS/MDS/SmartEvent IPs
let traversal_patterns = dynamic(["../", "..%2f", "%2e%2e", "%252e%252e", "..;/", "..%5c"]);
CommonSecurityLog
| where TimeGenerated > ago(lookback)
| where DestinationIP in (mgmt_ips) or DeviceVendor =~ "Check Point"
| where RequestURL has_any (traversal_patterns) or Message has_any (traversal_patterns)
| extend IsScriptUpload = (RequestMethod in ("POST", "PUT") and (RequestURL has_any (".jsp", ".php", ".sh", "upload")))
| summarize FirstSeen=min(TimeGenerated), LastSeen=max(TimeGenerated), Requests=count(),
DistinctURLs=dcount(RequestURL), SampleURL=any(RequestURL)
by SourceIP, DestinationIP, RequestMethod, IsScriptUpload
| order by Requests desc;
// Pivot: identify ANY external source touching the management interface at all — it shouldn't happen
CommonSecurityLog
| where TimeGenerated > ago(lookback)
| where DestinationIP in (mgmt_ips) and DestinationPort in (443, 4434, 19009)
| where not(ipv4_is_private(SourceIP))
| summarize Connections=count(), FirstSeen=min(TimeGenerated), LastSeen=max(TimeGenerated)
by SourceIP, DestinationPort
| order by Connections desc;
-- Velociraptor artifact: CVE-2026-93616 post-exploitation triage on Gaia management servers
-- Deploy the Linux agent to SMS/MDS/Log Server/SmartEvent hosts (maintenance-window approved)
-- 1) Suspicious child processes of the web stack (webshell execution)
SELECT Pid, Ppid, Name, Exe, CommandLine, Username, CreateTime
FROM pslist()
WHERE (Ppid IN (SELECT Pid FROM pslist() WHERE Name =~ '(java|httpd|apache)'))
AND Name =~ '(sh|bash|python|perl|curl|wget|nc|ncat)'
-- 2) Recently created script files in web/upload/tmp directories (last 14 days)
SELECT FullPath, Size, Mtime, Ctime
FROM glob(globs=['/opt/*/web/**/*.jsp', '/opt/*/webapps/**/*.jsp',
'/var/log/**/uploads/**/*.sh', '/tmp/*.sh', '/tmp/*.py',
'/var/tmp/*.sh', '/var/tmp/*.py',
'/web/htdocs/**/*.php', '/web/htdocs/**/*.jsp'])
WHERE Mtime > now() - 1209600
ORDER BY Mtime DESC
-- 3) Established outbound connections from the management server (C2 beaconing)
SELECT Pid, Name, Laddr, Lport, Raddr, Rport, Status
FROM netstat()
WHERE Status =~ 'ESTABLISHED'
AND NOT Raddr =~ '^(10\\.|172\\.(1[6-9]|2[0-9]|3[01])\\.|192\\.168\\.|127\\.)'
#!/bin/bash
# CVE-2026-93616 — Gaia management server triage & hardening verification
# Run via expert mode on SMS / MDS / Log Server / SmartEvent. Read-only checks; review before executing any remediation lines.
echo "=== [1] Confirm installed Jumbo Hotfix / version (verify against Check Point advisory for CVE-2026-93616) ==="
clish -c "show installer status" 2>/dev/null || cpinfo -y all 2>/dev/null
echo "=== [2] Scan web-served and temp directories for recently modified script files (last 14 days) ==="
find /opt /web /var/log -type f \( -name "*.jsp" -o -name "*.jspx" -o -name "*.php" -o -name "*.war" \) -mtime -14 -ls 2>/dev/null | head -100
find /tmp /var/tmp -type f \( -name "*.sh" -o -name "*.py" -o -name "*.pl" \) -mtime -14 -ls 2>/dev/null | head -100
echo "=== [3] Grep management web access logs for traversal sequences ==="
grep -REl '(\.\./|\.\.%2f|%2e%2e|%252e%252e|\.\.;/)' /opt/*/web/logs/ /var/log/httpd* /var/log/apache* 2>/dev/null | while read -r f; do
echo "--- $f ---"; grep -E '(\.\./|\.\.%2f|%2e%2e|%252e%252e|\.\.;/)' "$f" | tail -20
done
echo "=== [4] Child processes of web stack (live check) ==="
for pid in $(pgrep -f 'java|httpd'); do ps --ppid "$pid" -o pid,ppid,user,comm,args 2>/dev/null; done | grep -E 'sh|bash|python|perl|curl|wget|nc' || echo "No suspicious children found."
echo "=== [5] Outbound established connections (C2 check) ==="
ss -tnp state established | grep -vE '10\.|172\.(1[6-9]|2[0-9]|3[01])\.|192\.168\.|127\.' || echo "No external established sessions."
echo "=== [6] Verify management interface is NOT internet-reachable (check NAT/ACL on upstream firewall separately) ==="
clish -c "show configuration interface" 2>/dev/null | grep -A2 -i 'management\|eth0'
echo "DONE — preserve all output and logs to forensic storage BEFORE patching, per CISA Forensics Triage Requirements."
Remediation
Immediate (next 24 hours):
- Confirm exposure. Determine whether your management interface is reachable from the internet or from untrusted internal segments. Check your perimeter NAT rules and firewall policy for any inbound rules targeting the SMS/MDS/SmartEvent hosts on web ports. If it is exposed and unpatched, treat it as potentially compromised — begin forensic triage before patching, per CISA's Forensics Triage Requirements referenced in the KEV entry.
- Isolate the management plane. Management interfaces should never be internet-facing. Restrict access to a dedicated, ACL'd management network or jump host. This is a durable control, not a workaround — do it regardless of patch status.
- Capture evidence. Snapshot the VM/appliance, export web access logs, process lists, established connections, and the file inventory from the scan above. Store off-box.
Short term (72 hours — align to BOD 26-04 KEV timelines):
- Apply the vendor mitigation. Follow Check Point's official advisory for CVE-2026-93616 via the Check Point Support Center — install the specified hotfix/Jumbo Hotfix take on every affected component: SMS, MDS (all domains), Log Servers, Multi-Domain Log Servers, and SmartEvent. Verify installation with
show installer statusand confirm the build against the advisory's fixed-version table. Do not assume the management server auto-updates — it does not. - Hunt retroactively. Run the Sigma/KQL/VQL content above across at least 30 days of retained logs. Traversal attempts are loud in access logs — if you see a successful traversal followed by a POST to an upload endpoint and then outbound connections, you are in incident-response mode, not patch management.
- Rotate credentials and SIC material if compromise is suspected. Gaia admin credentials, API keys, and Secure Internal Communication certificates on a compromised management server must all be considered burned. Re-establishing SIC to gateways is painful; skipping it after a confirmed compromise is worse.
If patching is not possible: Per BOD 26-04 and the KEV required action, discontinue use of the affected exposure path — in practical terms, air-gap the management interface behind an allowlist-only ACL permitting only your admin jump hosts, and compensate with the detections above while you schedule the upgrade. Unmitigated and internet-reachable is not an acceptable steady state.
Longer term:
- Enroll Check Point management logs (web access, Gaia audit, process telemetry via a Linux agent) into your SIEM permanently — most organizations have zero visibility into their management plane until an incident forces the issue.
- Add management-plane exposure to your external attack surface monitoring. Your SMS should never appear in Shodan/Censys results.
- Bake KEV-driven SLAs into your vulnerability management program: KEV additions get a 72-hour remediation clock, full stop.
The KEV catalog exists because these vulnerabilities are being exploited against real organizations by real adversaries. CVE-2026-93616 targets the system that watches everything else. Patch it, hunt it, and verify the blast radius is zero.
Related Resources
Security Arsenal Penetration Testing Services AlertMonitor Platform Book a SOC Assessment vulnerability-management Intel Hub
Is your security operations ready?
Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.