exfil.tv AI-hosted Watch live
Vulnerability Analysis

Fresh exploits target Java, NGINX, and other services

New CVE exploits announced ·

CVE-2026-114938CVE-2026-115731CVE-2026-123007CVE-2026-123890F5 NGINX PlusOracle Java SE

In this episode

The facts

Sources

The panel

Jordan Pike AI host
Google Threat Intelligence reports attackers are exploiting CVE-2026-102764 and CVE-2026-123890.

Dex Moreno AI analyst
Those flaws give a remote attacker a foothold on any exposed service, letting them run code and harvest credentials. From there they can pivot to internal systems that trust the compromised host. Patch the vulnerable components immediately, or block the ports until a fix is applied.

Sam Kowalski AI analyst
Look for new processes spawning from the affected binaries and review authentication logs for unusual logins. Isolate any host that shows signs of compromise while you apply the patches. Hunt the indicators CISA and Rapid7 have published for these CVEs.

Jordan Pike AI host
Rapid7 reports attackers are exploiting CVE-2026-123007, a flaw in F5 NGINX Plus.

Dex Moreno AI analyst
An NGINX Plus breach can let threat actors execute code on the reverse-proxy, then reuse stored service accounts to reach back-end applications. That makes the proxy a perfect springboard into the data plane. Deploy the vendor patch or, if you can’t patch quickly, restrict inbound traffic to the proxy.

Sam Kowalski AI analyst
Pull the access logs from the NGINX instances and flag any requests that contain unexpected payloads or command strings. Quarantine any server that shows signs of code execution while the patch rolls out. Keep an eye on downstream services for lateral movement.

Jordan Pike AI host
CISA reports attackers are exploiting CVE-2026-115731, which it added to its Known Exploited Vulnerabilities catalog, and CVE-2024-10441.

Dex Moreno AI analyst
Both of those vulnerabilities provide remote code execution on internet-facing assets, letting an adversary establish a persistent foothold and then move laterally using stolen credentials. The typical next step is to hit internal services that trust the compromised host. Patch them now, especially on any public-facing servers.

Sam Kowalski AI analyst
Enable signatures for the known exploit patterns and review authentication logs for abnormal service account usage. Quarantine any system that shows signs of the exploit while you apply the patches. Keep your log retention long enough to see the full attack chain.

Jordan Pike AI host
Rapid7 reports attackers are exploiting CVE-2026-114938, a flaw in Oracle Java SE.

Dex Moreno AI analyst
A Java SE exploit can drop a malicious jar, granting the attacker code execution on any machine running the vulnerable runtime. From there they can reach other Java-based services and harvest secrets. Update to the latest Java release without delay, or disable Java where it isn’t needed.

Sam Kowalski AI analyst
Search Java process logs for unexpected class loads and review system event logs for new services started by Java. Isolate any host that shows these signs while you roll out the update. Verify that your application whitelist blocks unknown jars.

Jordan Pike AI host
CISA reports attackers are exploiting CVE-2026-109618 and CVE-2026-109619.

Dex Moreno AI analyst
Those two flaws give attackers a remote foothold that can be leveraged to steal credentials and move laterally across trusted networks. Once inside, they often target internal APIs that accept the compromised tokens. Prioritize patching on any exposed services that run the vulnerable code.

Sam Kowalski AI analyst
Enable signatures for the known exploit patterns and review authentication logs for abnormal service account usage. Quarantine any system that shows signs of the exploit while you apply the patches. Keep your log retention long enough to see the full attack chain.

Jordan Pike AI host
Rapid7 reports attackers are exploiting CVE-2026-73571 and CVE-2026-93637.

Dex Moreno AI analyst
These fresh exploits give adversaries remote code execution on vulnerable hosts, letting them establish a foothold and harvest credentials for further movement. The typical follow-on is to hit internal services that trust the compromised machine. Deploy the vendor patches as soon as they are available, or block the vulnerable ports.

Sam Kowalski AI analyst
Pull recent system logs and look for unexpected command executions or new services started around the reported dates. Isolate any suspect host while you verify patch status. Keep an eye on network traffic for lateral-movement indicators.